Dataset Viewer
query
stringlengths 16
206
| hits
listlengths 20
20
|
|---|---|
Image-Guided Nanopositioning Scheme for SEM
|
[
{
"content": "Title: Contrast-Guided Image Interpolation Content: In this paper a contrast-guided image interpolation method is proposed that incorporates contrast information into the image interpolation process. Given the image under interpolation, four binary contrast-guided decision maps (CDMs) are generated and used to guide the interpolation filtering through two sequential stages: 1) the 45° and 135° CDMs for interpolating the diagonal pixels and 2) the 0° and 90° CDMs for interpolating the row and column pixels. After applying edge detection to the input image, the generation of a CDM lies in evaluating those nearby non-edge pixels of each detected edge for re-classifying them possibly as edge pixels. This decision is realized by solving two generalized diffusion equations over the computed directional variation (DV) fields using a derived numerical approach to diffuse or spread the contrast boundaries or edges, respectively. The amount of diffusion or spreading is proportional to the amount of local contrast measured at each detected edge. The diffused DV fields are then thresholded for yielding the binary CDMs, respectively. Therefore, the decision bands with variable widths will be created on each CDM. The two CDMs generated in each stage will be exploited as the guidance maps to conduct the interpolation process: for each declared edge pixel on the CDM, a 1-D directional filtering will be applied to estimate its associated to-be-interpolated pixel along the direction as indicated by the respective CDM; otherwise, a 2-D directionless or isotropic filtering will be used instead to estimate the associated missing pixels for each declared non-edge pixel. Extensive simulation results have clearly shown that the proposed contrast-guided image interpolation is superior to other state-of-the-art edge-guided image interpolation methods. In addition, the computational complexity is relatively low when compared with existing methods; hence, it is fairly attractive for real-time image applications.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "2d948d7f1187db65f9b0306327563bc0526d76d9",
"rank": 1,
"score": 108857
},
{
"content": "Title: Guided Image Filtering Content: In this paper, we propose a novel explicit image filter called guided filter. Derived from a local linear model, the guided filter computes the filtering output by considering the content of a guidance image, which can be the input image itself or another different image. The guided filter can be used as an edge-preserving smoothing operator like the popular bilateral filter [1], but it has better behaviors near edges. The guided filter is also a more generic concept beyond smoothing: It can transfer the structures of the guidance image to the filtering output, enabling new filtering applications like dehazing and guided feathering. Moreover, the guided filter naturally has a fast and nonapproximate linear time algorithm, regardless of the kernel size and the intensity range. Currently, it is one of the fastest edge-preserving filters. Experiments show that the guided filter is both effective and efficient in a great variety of computer vision and computer graphics applications, including edge-aware smoothing, detail enhancement, HDR compression, image matting/feathering, dehazing, joint upsampling, etc.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "01dc8e05ad9590b3a1bf2f42226123c7da4b9fd1",
"rank": 2,
"score": 103186
},
{
"content": "Title: Design, Analysis, and Test of a Novel 2-DOF Nanopositioning System Driven by Dual Mode Content: Piezodriven flexure-based motion stages, with a large workspace and high positioning precision, are really attractive for the realization of high-performance atomic force microscope (AFM) scanning. In this paper, a modified lever displacement amplifier is proposed for the mechanism design of a novel compliant two-degree-of-freedom (2-DOF) nanopositioning stage, which can be selected to drive in dual modes. Besides, the modified double four-bar parallelogram, P (P denotes prismatic) joints are adopted in designing the flexure limbs. The established models for the mechanical performance evaluation of the stage, in terms of kinetostatics, dynamics, and workspace, are validated by the finite-element analysis. After a series of dimension optimizations carried out through the particle swarm optimization algorithm, a novel active disturbance rejection controller, including the nonlinearity tracking differentiator, the extended state observer, and the nonlinear state error feedback, is proposed to automatically estimate and suppress plant uncertainties arising from the hysteresis nonlinearity, creep effect, sensor noises, and unknown disturbances. The simulation and prototype test results indicate that the first natural frequency of the proposed stage is approximated to be 831 Hz, the amplification ratio in two axes is about 4.2, and the workspace is 119.7 μm × 121.4 μm, while the cross coupling between the two axes is kept within 2%. All the results prove that the developed stage possesses a good property for high-performance AFM scanning.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "ec28e43d33c838921d7e68fc7fd6a0cd521336a3",
"rank": 3,
"score": 101501
},
{
"content": "Title: Reverse Engineering Flash EEPROM Memories Using Scanning Electron Microscopy Content: In this article, a methodology to extract Flash EEPROM memory contents is presented. Samples are first backside prepared to expose the tunnel oxide of floating gate transistors. Then, a Scanning Electron Microscope (SEM) in the so called Passive Voltage Contrast (PVC) mode allows distinguishing ‘0’ and ‘1’ bit values stored in individual memory cell. Using SEM operator-free acquisition and standard image processing technique we demonstrate the possible automating of such technique over a full memory. The presented fast, efficient and low cost technique is successfully implemented on 0.35μm technology node microcontrollers and on a 0.21μm smart card type integrated circuit. The technique is at least two orders of magnitude faster than state-of-the-art Scanning Probe Microscopy (SPM) methods. Without adequate protection an adversary could obtain the full memory array content within minutes. The technique is a first step for reverse engineering secure embedded systems.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "802b80852996d87dc16082b86f6e77115eb6c9a6",
"rank": 4,
"score": 95082
},
{
"content": "Title: Image-guided full waveform inversion Content: Multiple problems, including high computational cost, spurious local minima, and solutions with no geologic sense, have prevented widespread application of full waveform inversion (FWI), especially FWI of seismic reflections. These problems are fundamentally related to a large number of model parameters and to the absence of low frequencies in recorded seismograms. Instead of inverting for all the parameters in a dense model, image-guided full waveform inversion inverts for a sparse model space that contains far fewer parameters. We represent a model with a sparse set of values, and from these values, we use image-guided interpolation (IGI) and its adjoint operator to compute finelyand uniformly-sampled models that can fit recorded data in FWI. Because of this sparse representation, image-guided FWI updates more blocky models, and this blockiness in the model space mitigates the absence of low frequencies in recorded data. Moreover, IGI honors imaged structures, so image-guided FWI built in this way yields models that are geologically sensible.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "f7686113611ac3d03c1cd412ebf46ba5e35b3071",
"rank": 5,
"score": 90476
},
{
"content": "Title: A novel explicit multi-focus image fusion method Content: We propose a multi-focus image fusion method. The energy of Laplacian of input images is obtained to decide which portions of the input images are in better focus, and the majority filter offers a means of spreading the focused regions to neighborhood. The edge-preserving guided image filter is applied to reduce the block effect. We conduct several experiments which demonstrate that it provided a better performance than the state of the art fusion methods in visual and quantitative evaluations.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "f30db882c88867e88c4f757568efb10abb069cd7",
"rank": 6,
"score": 89917
},
{
"content": "Title: Filtering of Video using Guided Filter Content: Image processing is one of the most important areas of signal processing. Image processing is applicable for the improvement of pictorial improvement for human perception and autonomous machine application also the efficient storage and transmission. In this paper guided filter is proposed which can be applicable to both images as well as to video. This guided filter has many advantages over the previous version of filter like mean filter, median filter, sobel filter, bilateral filter, etc. Among of those bilateral filter is more popular. But this bilateral filter has certain limitations like gradient reversal effect this introduces false edges in the image. Proposed guided filter can smooth image/video and also preserved the edge that is information at the edges of image/video but this filter has better behaviors near edges. Also the fast implementation of guided filter is possible. This filter also has algorithm implementation regardless of kernel size. Guided filter gives better output than that of bilateral filter after filtering this bilateral filter gives output like cartoon this also a problem with this. The guided filter is both effective and efficient in many applications of computer vision and computer graphics, etc. Filtering techniques are used in image and video processing applications for various technologies.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "8144c5e362da8dff8def2c8cba91bedca9c55056",
"rank": 7,
"score": 89316
},
{
"content": "Title: Guided filtering for PRNU-based localization of small-size image forgeries Content: PRNU-based techniques guarantee a good forgery detection performance irrespective of the specific type of forgery. The presence or absence of the camera PRNU pattern is detected by a correlation test. Given the very low power of the PRNU signal, however, the correlation must be averaged over a pretty large window, reducing the algorithm's ability to reveal small forgeries. To improve resolution, we estimate correlation with a spatially adaptive filtering technique, with weights computed over a suitable pilot image. Implementation efficiency is achieved by resorting to the recently proposed guided filters. Experiments prove that the proposed filtering strategy allows for a much better detection performance in the case of small forgeries.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "4ee0ad8e256523256c8d21790189388ed4beca7e",
"rank": 8,
"score": 89281
},
{
"content": "Title: Gradient Domain Guided Image Filtering Content: Guided image filter (GIF) is a well-known local filter for its edge-preserving property and low computational complexity. Unfortunately, the GIF may suffer from halo artifacts, because the local linear model used in the GIF cannot represent the image well near some edges. In this paper, a gradient domain GIF is proposed by incorporating an explicit first-order edge-aware constraint. The edge-aware constraint makes edges be preserved better. To illustrate the efficiency of the proposed filter, the proposed gradient domain GIF is applied for single-image detail enhancement, tone mapping of high dynamic range images and image saliency detection. Both theoretical analysis and experimental results prove that the proposed gradient domain GIF can produce better resultant images, especially near the edges, where halos appear in the original GIF.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "e1257afedb5699bd1c76c6e5cd9b8281d72245c6",
"rank": 9,
"score": 83166
},
{
"content": "Title: Learning Dynamic Guidance for Depth Image Enhancement Content: The depth images acquired by consumer depth sensors (e.g., Kinect and ToF) usually are of low resolution and insufficient quality. One natural solution is to incorporate with high resolution RGB camera for exploiting their statistical correlation. However, most existing methods are intuitive and limited in characterizing the complex and dynamic dependency between intensity and depth images. To address these limitations, we propose a weighted analysis representation model for guided depth image enhancement, which advances the conventional methods in two aspects: (i) task driven learning and (ii) dynamic guidance. First, we generalize the analysis representation model by including a guided weight function for dependency modeling. And the task-driven learning formulation is introduced to obtain the optimized guidance tailored to specific enhancement task. Second, the depth image is gradually enhanced along with the iterations, and thus the guidance should also be dynamically adjusted to account for the updating of depth image. To this end, stage-wise parameters are learned for dynamic guidance. Experiments on guided depth image upsampling and noisy depth image restoration validate the effectiveness of our method.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "29b09a84ff5f90aed578f75f4b41ae49e7d9b6d7",
"rank": 10,
"score": 82011
},
{
"content": "Title: Sharpness Enhancement and Super-Resolution of Around-View Monitor Images Content: In the wide-angle (WA) images embedded in an around-view monitor system, the subject(s) in the peripheral region is normally small and has little information. Furthermore, since the outer region suffers from the non-uniform blur phenomenon and artifact caused by the inherent optical characteristic of WA lenses, its visual quality tends to deteriorate. According to our experiments, conventional image enhancement techniques rarely improve the degraded visual quality of the outer region of WA images. In order to solve the above-mentioned problem, this paper proposes a joint sharpness enhancement (SE) and super-resolution (SR) algorithm which can improve the sharpness and resolution of WA images together. The proposed SE algorithm improves the sharpness of the deteriorated WA images by exploiting self-similarity. Also, the proposed SR algorithm generates super-resolved images by using high-resolution information which is classified according to the extended local binary pattern-based classifier and learned on a pattern basis. Experimental results show that the proposed scheme effectively improves the sharpness and resolution of the input deteriorated WA images. Even in terms of quantitative metrics such as just noticeable blur, structural similarity, and peak signal-to-noise ratio. Finally, the proposed scheme guarantees real-time processing such that it achieves 720p video at 29 Hz on a low cost GPU platform.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "64aa6e5e7b3f8a29bad2e97463ffb7bd43a8af9d",
"rank": 11,
"score": 80446
},
{
"content": "Title: A Virtual Feedback Assistance System for Remote Operation of a 3DOF Micromanipulator in Micro-Nanorobotic Manipulation Content: Manipulation in micro or nanoscale with robotic manipulators under observation of electron microscopes is a widely used strategy for fabrication of nanodevices and nanoscale material property characterization. These types of manipulation systems can handle the relatively larger scale of objects. However, the complexity of manipulation increases highly for 3D manipulation. Since the manipulation system consists of multiple components including manipulator, microscope, and also some end-effector tools, a proper offline visualization of the system is necessary for operation. Therefore, we propose a web-based virtual interface between the user and the actual manipulator operated under digital microscope initially. It gives the operator 3D positional feedback from the virtual model by mapping data read during remote operation. The same interface is used for remote operation of the manipulator within the SEM chamber and a manipulation task is performed.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "a0fd813b9218813e1b020d03a3099de7677dd145",
"rank": 12,
"score": 79755
},
{
"content": "Title: Fast End-to-End Trainable Guided Filter Content: Image processing and pixel-wise dense prediction have been advanced by harnessing the capabilities of deep learning. One central issue of deep learning is the limited capacity to handle joint upsampling. We present a deep learning building block for joint upsampling, namely guided filtering layer. This layer aims at efficiently generating the high-resolution output given the corresponding low-resolution one and a high-resolution guidance map. The proposed layer is composed of a guided filter, which is reformulated as a fully differentiable block. To this end, we show that a guided filter can be expressed as a group of spatial varying linear transformation matrices. This layer could be integrated with the convolutional neural networks (CNNs) and jointly optimized through end-to-end training. To further take advantage of end-to-end training, we plug in a trainable transformation function that generates task-specific guidance maps. By integrating the CNNs and the proposed layer, we form deep guided filtering networks. The proposed networks are evaluated on five advanced image processing tasks. Experiments on MIT-Adobe FiveK Dataset demonstrate that the proposed approach runs 10-100× faster and achieves the state-of-the-art performance. We also show that the proposed guided filtering layer helps to improve the performance of multiple pixel-wise dense prediction tasks. The code is available at https://github.com/wuhuikai/DeepGuidedFilter.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "68e07fdfa8de4eb7ace6374ce9c16168317414e3",
"rank": 13,
"score": 78998
},
{
"content": "Title: Image detail enhancement with spatially guided filters Content: In recent years, enhancing image details without introducing artifacts has been attracting much attention in image processing community. Various image filtering methods have been proposed to achieve this goal. However, existing methods usually treat all pixels equally during the filtering process without considering the relationship between filtering strengths and image contents. In this paper, we address this issue by spatially distributing the filtering strengths with simple low-level features. First, to determine the pixel-wise filtering strength, we construct a spatially guided map, which exploits the spatial influence of image details based on the edge response of an image. Then, we further integrate this guided map into two state-of-the-art image filters and apply the improved filters to the task of image detail enhancement. Experiments demonstrate that our results generate better content-specific visual effects and introduce much less artifacts.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "7085851515bf3e7a5ce5f5e0cb44d4559a3c9819",
"rank": 14,
"score": 77197
},
{
"content": "Title: JOINT IMAGE SHARPENING AND DENOISING BY 3 D TRANSFORM-DOMAIN COLLABORATIVE FILTERING Content: In order to simultaneously sharpen image details and attenuate noise, we propose to combine the recent blockmatching and 3D Þltering (BM3D) denoising approach, based on 3D transform-domain collaborative Þltering, with alpha-rooting, a transform-domain sharpening technique. The BM3D exploits grouping of similar image blocks into 3D arrays (groups) on which collaborative Þltering (by hard-thresholding) is applied. We propose two approaches of sharpening by alpha-rooting; the Þrst applies alpharooting individually on the 2D transform spectra of each grouped block; the second applies alpha-rooting on the 3D-transform spectra of each 3D array in order to sharpen Þne image details shared by all grouped blocks and further enhance the interblock differences. The conducted experiments with the proposed method show that it can preserve and sharpen Þne image details and effectively attenuate noise.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "4fd69173cabb3d4377432d70488938ac533a5ac3",
"rank": 15,
"score": 76970
},
{
"content": "Title: Segmentation Guided Attention Networks for Visual Question Answering Content: In this paper we propose to solve the problem of Visual Question Answering by using a novel segmentation guided attention based network which we call SegAttendNet. We use image segmentation maps, generated by a Fully Convolutional Deep Neural Network to refine our attention maps and use these refined attention maps to make the model focus on the relevant parts of the image to answer a question. The refined attention maps are used by the LSTM network to learn to produce the answer. We presently train our model on the visual7W dataset and do a category wise evaluation of the 7 question categories. We achieve state of the art results on this dataset and beat the previous benchmark on this dataset by a 1.5% margin improving the question answering accuracy from 54.1% to 55.6% and demonstrate improvements in each of the question categories. We also visualize our generated attention maps and note their improvement over the attention maps generated by the previous best approach.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "e8691980eeb827b10cdfb4cc402b3f43f020bc6a",
"rank": 16,
"score": 75216
},
{
"content": "Title: A Novel Fuzzy C-Means Clustering Algorithm for Image Thresholding Content: Image thresholding has played an important role in image segmentation. In this paper, we present a novel spatially weighted fuzzy c-means (SWFCM) clustering algorithm for image thresholding. The algorithm is formulated by incorporating the spatial neighborhood information into the standard FCM clustering algorithm. Two improved implementations of the k-nearest neighbor (k-NN) algorithm are introduced for calculating the weight in the SWFCM algorithm so as to improve the performance of image thresholding. To speed up the FCM algorithm, the iteration is carried out with the gray level histogram of image instead of the conventional whole data of image. The performance of the algorithm is compared with those of an existed fuzzy thresholding algorithm and widely applied between variance and entropy methods. Experimental results with synthetic and real images indicate the proposed approach is effective and efficient. In addition, due to the neighborhood model, the proposed method is more tolerant to noise.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "ebbef7e8b7a1ea133999561ab279e51b961d31ec",
"rank": 17,
"score": 74872
},
{
"content": "Title: PLANT LEAF RECOGNITION Content: This paper proposes an automated system for recognizing plant species based on leaf images. Plant leaf images corresponding to three plant types, are analyzed using three different shape modelling techniques, the first two based on the Moments-Invariant (M-I) model and the Centroid-Radii (C-R) model and the third based on a proposed technique of Binary-Superposition (B-S). For the M-I model the first four central normalized moments have been considered. For the C-R model an edge detector has been used to identify the boundary of the leaf shape and 36 radii at 10 degree angular separation have been used to build the shape vector. The proposed approach consists of comparing binary versions of the leaf images through superposition and using the sum of non-zero pixel values of the resultant as the feature vector. The data set for experimentations consists of 180 images divided into training and testing sets and comparison between them is done using Manhattan, Euclidean and intersection norms. Accuracies obtained using the proposed technique is seen to be an improvement over the M-I and C-R based techniques, and comparable to the best figures reported in extant literature.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "ddf1f27694928a729aefa6ffd6faabfd8ebf2842",
"rank": 18,
"score": 74548
},
{
"content": "Title: Blind detection of photomontage using higher order statistics Content: We investigate the prospect of using bicoherence features for blind image splicing detection. Image splicing is an essential operation for digital photomontaging, which in turn is a technique for creating image forgery. We examine the properties of bicoherence features on a data set, which contains image blocks of diverse image properties. We then demonstrate the limitation of the baseline bicoherence features for image splicing detection. Our investigation has led to two suggestions for improving the performance of bicoherence features, i.e., estimating the bicoherence features of the authentic counterpart and incorporating features that characterize the variance of the feature performance. The features derived from the suggestions are evaluated with support vector machine (SVM) classification and is shown to improve the image splicing detection accuracy from 62% to about 70%.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "2230381a078241c4385bb9c20b385e8f0da70b9b",
"rank": 19,
"score": 74005
},
{
"content": "Title: Corpus-Guided Sentence Generation of Natural Images Content: We propose a sentence generation strategy that describes images by predicting the most likely nouns, verbs, scenes and prepositions that make up the core sentence structure. The input are initial noisy estimates of the objects and scenes detected in the image using state of the art trained detectors. As predicting actions from still images directly is unreliable, we use a language model trained from the English Gigaword corpus to obtain their estimates; together with probabilities of co-located nouns, scenes and prepositions. We use these estimates as parameters on a HMM that models the sentence generation process, with hidden nodes as sentence components and image detections as the emissions. Experimental results show that our strategy of combining vision and language produces readable and descriptive sentences compared to naive strategies that use vision alone.",
"qid": "01273bd34dacfe9ef887b320f36934d2f9fa9b34",
"docid": "3a0a839012575ba455f2b84c2d043a35133285f9",
"rank": 20,
"score": 73859
}
] |
Efficient and secure data storage operations for mobile cloud computing
|
[
{
"content": "Title: Secure data processing framework for mobile cloud computing Content: In mobile cloud computing, mobile devices can rely on cloud computing and information storage resource to perform computationally intensive operations such as searching, data mining, and multimedia processing. In addition to providing traditional computation services, mobile cloud also enhances the operation of traditional ad hoc network by treating mobile devices as service nodes, e.g., sensing services. The sensed information, such as location coordinates, health related information, should be processed and stored in a secure fashion to protect user's privacy in the cloud. To this end, we present a new mobile cloud data processing framework through trust management and private data isolation. Finally, an implementation pilot for improving teenagers' driving safety, which is called FocusDrive, is presented to demonstrate the solution.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "af77547dc79c67367675e76f28c3bbf3032a9a12",
"rank": 1,
"score": 185766
},
{
"content": "Title: Enabling Public Verifiability and Data Dynamics for Storage Security Content: Cloud Computing has been envisioned as the next-generation architecture of IT Enterprise. It moves the application software and databases to the centralized large data centers, where the management of the data and services may not be fully trustworthy. This unique paradigm brings about many new security challenges, which have not been well understood. This work studies the problem of ensuring the integrity of data storage in Cloud Computing. In particular, we consider the task of allowing a third party auditor (TPA), on behalf of the cloud client, to verify the integrity of the dynamic data stored in the cloud. The introduction of TPA eliminates the involvement of client through the auditing of whether his data stored in the cloud is indeed intact, which can be important in achieving economies of scale for Cloud Computing. The support for data dynamics via the most general forms of data operation, such as block modification, insertion and deletion, is also a significant step toward practicality, since services in Cloud Computing are not limited to archive or backup data only. While prior works on ensuring remote data integrity often lacks the support of either public verifiability or dynamic data operations, this paper achieves both. We first identify the difficulties and potential security problems of direct extensions with fully dynamic data updates from prior works and then show how to construct an elegant verification scheme for seamless integration of these two salient features in our protocol design. In particular, to achieve efficient data dynamics, we improve the Proof of Retrievability model [1] by manipulating the classic Merkle Hash Tree (MHT) construction for block tag authentication. Extensive security and performance analysis show that the proposed scheme is highly efficient and provably secure.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "9c22a8108c96c39ea930a5123ac1f18b6f4bbc60",
"rank": 2,
"score": 150395
},
{
"content": "Title: Help your mobile applications with fog computing Content: Cloud computing has paved a way for resource-constrained mobile devices to speed up their computing tasks and to expand their storage capacity. However, cloud computing is not necessary a panacea for all mobile applications. The high network latency to cloud data centers may not be ideal for delay-sensitive applications while storing everything on public clouds risks users' security and privacy. In this paper, we discuss two preliminary ideas, one for mobile application offloading and the other for mobile storage expansion, by leveraging the edge intelligence offered by fog computing to help mobile applications. Preliminary experiments conducted based on implemented prototypes show that fog computing can provide an effective and sometimes better alternative to help mobile applications.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "541c3ab3ce75594c403126413b9c866fa7fba57a",
"rank": 3,
"score": 149198
},
{
"content": "Title: An Optimization Approach for Utilizing Cloud Services for Mobile Devices in Cloud Environment Content: Abstract. Mobile cloud computing has emerged aiming at assisting mobile devices in processing computationally or data intensive tasks using cloud resources. This paper presents an optimization approach for utilizing cloud services for mobile client in mobile cloud, which considers the benefit of both mobile device users and cloud datacenters. The mobile cloud service provisioning optimization is conducted in parallel under the deadline, budget and energy expenditure constraint. Mobile cloud provider runs multiple VMs to execute the jobs for mobile device users, the cloud providers want to maximize the revenue and minimize the electrical cost. The mobile device user gives the suitable payment to the cloud datacenter provider for available cloud resources for optimize the benefit. The paper proposes a distributed optimization algorithm for utilizing cloud services for mobile devices. The experiment is to test convergence of the proposed algorithm and also compare it with other related work. The experiments study the impacts of job arrival rate, deadline and mobility speeds on energy consumption ratio, execution success ratio, resource allocation efficiency and cost. The experiment shows that the proposed algorithm outperforms other related work in terms of some performance metrics such as allocation efficiency.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "de33698f7b2264bf7313f43d1de8c2d19e2a2f7a",
"rank": 4,
"score": 142749
},
{
"content": "Title: A Privacy-Aware Authentication Scheme for Distributed Mobile Cloud Computing Services Content: In modern societies, the number of mobile users has dramatically risen in recent years. In this paper, an efficient authentication scheme for distributed mobile cloud computing services is proposed. The proposed scheme provides security and convenience for mobile users to access multiple mobile cloud computing services from multiple service providers using only a single private key. The security strength of the proposed scheme is based on bilinear pairing cryptosystem and dynamic nonce generation. In addition, the scheme supports mutual authentication, key exchange, user anonymity, and user untraceability. From system implementation point of view, verification tables are not required for the trusted smart card generator (SCG) service and cloud computing service providers when adopting the proposed scheme. In consequence, this scheme reduces the usage of memory spaces on these corresponding service providers. In one mobile user authentication session, only the targeted cloud service provider needs to interact with the service requestor (user). The trusted SCG serves as the secure key distributor for distributed cloud service providers and mobile clients. In the proposed scheme, the trusted SCG service is not involved in individual user authentication process. With this design, our scheme reduces authentication processing time required by communication and computation between cloud service providers and traditional trusted third party service. Formal security proof and performance analyses are conducted to show that the scheme is both secure and efficient.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "43d40cf9251ba497f6ea3957bfc3c189fd11d421",
"rank": 5,
"score": 141812
},
{
"content": "Title: DCnet: A new data center network architecture Content: The migration of computation and storage to cloud data centers has imposed new communication requirements and a need for fundamental changes in infrastructure. The cloud paradigm means that significant network communication will occur between servers in data centers. Furthermore, the migration of virtual machines and containers across physical servers means that in addition to traditional requirements of efficiency, scalability, and easy manageability, future data center networks must also support efficient mobility. To solve the problem, the paper proposes a radical redesign for data center networks, using a new architectural approach that we call DCnet. The DCnet approach changes the addressing and routing schemes at layers 2 and 3 completely. The new addressing mechanism allows an organization to assign addresses that span multiple data centers and permits VM mobility across data centers. Despite radical changes in addressing and routing, the DCnet architecture retains basics from Ethernet and IP, allowing existing hardware building blocks to be reused. The paper uses IPv6 in examples, but the new addressing model and address mobility approach can be applied to IPv4. By assigning each addressable entity a unique ID, and using the ID as the host suffix in an IP address, DCnet allows legacy operating systems and applications to be used without changes. The paper reports a simulation study that uses mininet and SDN to demonstrate the feasibility of the DCnet approach.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "9acb353a94c258584c9048e8a700dbed1008d64a",
"rank": 6,
"score": 140453
},
{
"content": "Title: Privacy preservation and public auditing for cloud data using ASS in Multi-cloud Content: Cloud computing is the new business paradigm introduced in the IT market in the current decade providing the customers with the option to pay for the amount of service utilized. The use of cloud computing has increased rapidly in many organizations. Ensuring security of data in a cloud is a major factor in the cloud computing environment, as users often store fragile information in cloud storage. Public auditing on shared data stored in the cloud is done by exploiting Aggregate Signature Scheme. With this mechanism, the identity of the signer of various blocks of shared data is kept private from public verifiers, who are able to verify efficiently the integrity of shared data without retrieving the entire file. However traceability is a major issue. A Multi-clouds Database Model (MCDB) which is based on Multi-clouds service providers instead of using single cloud service provider such as in Amazon cloud service is used to implement traceability. Also, dealing with single cloud providers is predicted to become less popular with customers due to risks of service availability failure and the possibility of malicious insiders in the single cloud. A movement towards multi-clouds is highly efficient and this system would also check for the data's which are attacked by malicious users, with short time and cost requirement.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "9c79d6f2ce0e14a2e6d9d574cdc185fa204bfae7",
"rank": 7,
"score": 137258
},
{
"content": "Title: Towards Automated Cost-Efficient Data Management for Federated Cloud Services Content: Cloud computing has transformed the accessibility and usage of information technology resources, by offering them as services via the Internet. Cloud service provisioning spans across infrastructure, platform and software levels. The management of the services at these levels is based on monitoring. The analysis of monitoring data provides insight into Cloud operations in order to make informed decisions. Due to the emergence of numerous heterogenous Cloud platforms with proprietary APIs, service and monitoring data are being formatted using diverse and mostly incompatible data interchange formats. This results to interoperability issues and makes the analysis of monitoring data from multi-Cloud service deployments difficult to handle. The existing research efforts on data interchange formats have been mainly focused on general performance analyses. Little or no effort has been channelled towards a combination of multiple data interchange formats based on data type to achieve efficient serialisation that can facilitate interoperability in federated Clouds, and also reduce the size of data and bandwidth utilisation cost. This paper addresses these issues by presenting automated framework that is capable of automatically selecting the most suitable data interchange formats for achieving an efficient formatting and serialisation outcome. The goal of the framework is to enable robust and transparent communication within and between multiple Cloud deployments. Based on three use case scenarios, we evaluate the proposed framework to demonstrate its efficacy in formatting and serialising data.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "cadefcf12a0ae14baff8a1fa67819b4e4f644396",
"rank": 8,
"score": 136221
},
{
"content": "Title: Security and Privacy in Cloud Computing: Vision, Trends, and Challenges Content: Cloud computing offers cost-effective solutions via a variety of flexible services. However, security concerns related to managing data, applications, and interactions hamper the rapid deployment of cloud-based services on a large scale. Although many solutions exist, efficiency, scalability, and provable security still have issues that need to be properly addressed. This article explores the various challenges, existing solutions, and limitations of cloud security, with a focus on data utilization management aspects, including data storage, data analytics, and access control. The article concludes with a discussion on future research directions that might lead to more trustworthy cloud security and privacy.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "d40355618f3b647584770659d9ba1707b233aa76",
"rank": 9,
"score": 134074
},
{
"content": "Title: Twin Clouds: An Architecture for Secure Cloud Computing (Extended Abstract) Content: Cloud computing promises a more cost effective enabling technology to outsource storage and computations. Existing approaches for secure outsourcing of data and arbitrary computations are either based on a single tamper-proof hardware, or based on recently proposed fully homomorphic encryption. The hardware based solutions are not scaleable, and fully homomorphic encryption is currently only of theoretical interest and very inefficient. In this paper we propose an architecture for secure outsourcing of data and arbitrary computations to an untrusted commodity cloud. In our approach, the user communicates with a trusted cloud (either a private cloud or built from multiple secure hardware modules) which encrypts and verifies the data stored and operations performed in the untrusted commodity cloud. We split the computations such that the trusted cloud is mostly used for security-critical operations in the less time-critical setup phase, whereas queries to the outsourced data are processed in parallel by the fast commodity cloud on encrypted data.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "3ffdcfdd7511beab79539a96c2ae805c40160a36",
"rank": 10,
"score": 133506
},
{
"content": "Title: A framework based on RSA and AES encryption algorithms for cloud computing services Content: Cloud computing is an emerging computing model in which resources of the computing communications are provided as services over the Internet. Privacy and security of cloud storage services are very important and become a challenge in cloud computing due to loss of control over data and its dependence on the cloud computing provider. While there is a huge amount of transferring data in cloud system, the risk of accessing data by attackers raises. Considering the problem of building a secure cloud storage service, current scheme is proposed which is based on combination of RSA and AES encryption methods to share the data among users in a secure cloud system. The proposed method allows providing difficulty for attackers as well as reducing the time of information transmission between user and cloud data storage.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "4c9774c5e57a4b7535eb19f6584f75c8b9c2cdcc",
"rank": 11,
"score": 133341
},
{
"content": "Title: Ensuring Data Integrity in Cloud Computing Content: Cloud computing provides convenient on-demand network access to a shared pool of configurable computing resources. The resources can be rapidly deployed with great efficiency and minimal management overhead. Cloud is an insecure computing platform from the view point of the cloud users, the system must design mechanisms that not only protect sensitive information by enabling computations with encrypted data, but also protect users from malicious behaviours by enabling the validation of the computation result. In this paper, we propose a new data encoding scheme called layered interleaving, designed for time-sensitive packet recovery in the presence of bursty loss. It is high-speed data recovery scheme with minimal loss probability and using a forward error correction scheme to handle bursty loss. The proposed approach is highly efficient in recovering the singleton losses almost immediately and from bursty data losses.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "f9823bc7eec44a9a6cd7b629c8f6430fe82877fd",
"rank": 12,
"score": 133107
},
{
"content": "Title: DAC-MACS: Effective Data Access Control for Multiauthority Cloud Storage Systems Content: Data access control is an effective way to ensure data security in the cloud. However, due to data outsourcing and untrusted cloud servers, the data access control becomes a challenging issue in cloud storage systems. Existing access control schemes are no longer applicable to cloud storage systems, because they either produce multiple encrypted copies of the same data or require a fully trusted cloud server. Ciphertext-policy attribute-based encryption (CP-ABE) is a promising technique for access control of encrypted data. However, due to the inefficiency of decryption and revocation, existing CP-ABE schemes cannot be directly applied to construct a data access control scheme for multiauthority cloud storage systems, where users may hold attributes from multiple authorities. In this paper, we propose data access control for multiauthority cloud storage (DAC-MACS), an effective and secure data access control scheme with efficient decryption and revocation. Specifically, we construct a new multiauthority CP-ABE scheme with efficient decryption, and also design an efficient attribute revocation method that can achieve both forward security and backward security. We further propose an extensive data access control scheme (EDAC-MACS), which is secure under weaker security assumptions.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "4bb507d3d74354ecab6dad25f211bbb856bbb392",
"rank": 13,
"score": 132167
},
{
"content": "Title: Challenges of using homomorphic encryption to secure cloud computing Content: With the emergence of cloud computing, the concept of information security has become a major issue. Indeed, the security of such a system is the greatest concern of computer scientists, providers of cloud and organizations who want to adopt and benefit from these services. Cloud computing providers must implement concepts ensuring network security, hardware, data storage and strategies of control and access to services. All these elements help to preserve data security and ensuring the availability of services associated with the Cloud, to better satisfy clients and acquire and build their trust. However, even if the data storage security in cloud servers is assured, reluctance remain when it comes to process the confidential data. Indeed, the fear that sensitive data is being used is a major obstacle in the adoption of cloud services by enterprises. To overcome this obstacle, the use of methods that can perform operations on encrypted data without knowing the secret key, seems to be an effective way to strengthen the confidentiality of information. In this paper we will examine the challenges facing Homomorphic Encryption methods to allow suppliers of cloud to perform operations on encrypted data, and provide the same results after treatment, as if they were performing calculations on raw data.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "7da05dc8c992126f09610d22415d3a6f8bee324a",
"rank": 14,
"score": 131794
},
{
"content": "Title: Opportunistic Computation Offloading in Mobile Edge Cloud Computing Environments Content: The dynamic mobility and limitations in computational power, battery resources, and memory availability are main bottlenecks in fully harnessing mobile devices as data mining platforms. Therefore, the mobile devices are augmented with cloud resources in mobile edge cloud computing (MECC) environments to seamlessly execute data mining tasks. The MECC infrastructures provide compute, network, and storage services in one-hop wireless distance from mobile devices to minimize the latency in communication as well as provide localized computations to reduce the burden on federated cloud systems. However, when and how to offload the computation is a hard problem. In this paper, we present an opportunistic computation offloading scheme to efficiently execute data mining tasks in MECC environments. The scheme provides the suitable execution mode after analyzing the amount of unprocessed data, privacy configurations, contextual information, and available on-board local resources (memory, CPU, and battery power). We develop a mobile application for online activity recognition and evaluate the proposed scheme using the event data stream of 5 million activities collected from 12 users for 15 days. The experiments show significant improvement in execution time and battery power consumption resulting in 98% data reduction.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "ec1df457a2be681227f79de3ce932fccb65ee2bb",
"rank": 15,
"score": 131297
},
{
"content": "Title: ScienceDirect The 2 nd International Workshop on Communications and Sensor Networks ( ComSense-2014 ) A cloud-based healthcare framework for security and patients ' data privacy using wireless body area networks Content: The recent developments in remote healthcare systems have witnessed significant interests from IT industry (Microsoft, Google, VMware etc) that provide ubiquitous and easily deployable healthcare systems. These systems provide a platform to share medical information, applications, and infrastructure in a ubiquitous and fully automated manner. Communication security and patients' data privacy are the aspects that would increase the confidence of users in such remote healthcare systems. This paper presents a secure cloud-based mobile healthcare framework using wireless body area networks (WBANs). The research work presented here is twofold: first, it attempts to secure the inter-sensor communication by multi-biometric based key generation scheme in WBANs; and secondly, the electronic medical records (EMRs) are securely stored in the hospital community cloud and privacy of the patients' data is preserved. The evaluation and analysis shows that the proposed multi-biometric based mechanism provides significant security measures due to its highly efficient key generation mechanism.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "8eb3a7f114fbfcd7708dda984e0378376ced8843",
"rank": 16,
"score": 128834
},
{
"content": "Title: P-MOD: Secure Privilege-Based Multilevel Organizational Data-Sharing in Cloud Computing Content: Cloud computing has changed the way enterprises store, access and share data. Data is constantly being uploaded to the cloud and shared within an organization built on a hierarchy of many different individuals that are given certain data access privileges. With more data storage needs turning over to the cloud, finding a secure and efficient data access structure has become a major research issue. With different access privileges, individuals with more privileges (at higher levels of the hierarchy) are granted access to more sensitive data than those with fewer privileges (at lower levels of the hierarchy). In this paper, a Privilege-based Multilevel Organizational Data-sharing scheme (P-MOD) is proposed that incorporates a privilege-based access structure into an attribute-based encryption mechanism to handle these concerns. Each level of the privilege-based access structure is affiliated with an access policy that is uniquely defined by specific attributes. Data is then encrypted under each access policy at every level to grant access to specific data users based on their data access privileges. An individual ranked at a certain level can decrypt the ciphertext (at that specific level) if and only if that individual owns a correct set of attributes that can satisfy the access policy of that level. The user may also decrypt the ciphertexts at the lower levels with respect to the user’s level. Security analysis shows that P-MOD is secure against adaptively chosen plaintext attack assuming the DBDH assumption holds. The comprehensive performance analysis demonstrates that PMOD is more efficient in computational complexity and storage space than the existing schemes in secure data sharing within an organization.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "288eb3c71c76ef2a820c0c2e372436ce078224d0",
"rank": 17,
"score": 128754
},
{
"content": "Title: ESPAC: Enabling Security and Patient-centric Access Control for eHealth in cloud computing Content: We consider the problem of patient self-controlled access privilege to highly sensitive Personal Health Information (PHI), where PHI is expected to be securely stored in cloud storage for uninterrupted anytime, anywhere remote access. In order to assure the privacy of PHI, we propose Efficient and Secure Patient-centric Access Control (ESPAC) scheme which allows data requesters to have different access privileges based on their roles, and then assigns different attribute sets to them. Extensive security and performance analyses demonstrate that the ESPAC scheme is able to achieve desired security requirements with acceptable communication delay.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "f7cc9f54f910e7e491a1e62d04dd3430c3aa9316",
"rank": 18,
"score": 127571
},
{
"content": "Title: Towards secure mobile cloud computing: A survey Content: Mobile cloud computing is gaining popularity among mobile users. The ABI Research predicts that the number of mobile cloud computing subscribers is expected to grow from 42.8 million (1.1% of total mobile users) in 2008 to 998 million (19% of total mobile users) in 2014. Despite the hype achieved by mobile cloud computing, the growth of mobile cloud computing subscribers is still below expectations. According to the recent survey conducted by the International Data Corporation, most IT Executives and CEOs are not interested in adopting such services due to the risks associatedwith security and privacy. The security threats have become a hurdle in the rapid adaptability of the mobile cloud computing paradigm. Significant efforts have been devoted in research organizations and academia to build securemobile cloud computing environments and infrastructures. In spite of the efforts, there are a number of loopholes and challenges that still exist in the security policies of mobile cloud computing. This literature review: (a)highlights the current state of the artwork proposed to securemobile cloud computing infrastructures, (b) identifies the potential problems, and (c) provides a taxonomy of the state of the art. © 2012 Elsevier B.V. All rights reserved.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "70127f6cd77b92f0e7228179466edb4f259f56c2",
"rank": 19,
"score": 126717
},
{
"content": "Title: Secure Fine-Grained Access Control and Data Sharing for Dynamic Groups in the Cloud Content: Cloud computing is an emerging computing paradigm that enables users to store their data in a cloud server to enjoy scalable and on-demand services. Nevertheless, it also brings many security issues, since cloud service providers (CSPs) are not in the same trusted domain as users. To protect data privacy against untrusted CSPs, existing solutions apply cryptographic methods (e.g., encryption mechanisms) and provide decryption keys only to authorized users. However, sharing cloud data among authorized users at a fine-grained level is still a challenging issue, especially when dealing with dynamic user groups. In this paper, we propose a secure and efficient fine-grained access control and data sharing scheme for dynamic user groups by: 1) defining and enforcing access policies based on the attributes of the data; 2) permitting the key generation center to efficiently update user credentials for dynamic user groups; and 3) allowing some expensive computation tasks to be performed by untrusted CSPs without requiring any delegation key. Specifically, we first design an efficient revocable attribute-based encryption (ABE) scheme with the property of ciphertext delegation by exploiting and uniquely combining techniques of identity-based encryption, ABE, subset-cover framework, and ciphertext encoding mechanism. We then present a fine-grained access control and data sharing system for on-demand services with dynamic user groups in the cloud. The experimental data show that our proposed scheme is more efficient and scalable than the state-of-the-art solution.",
"qid": "012e396b02aa584cb74a65ae14af355e7c897858",
"docid": "be98ba27ffdaa99834d12a1aa9c905c7bc6848c1",
"rank": 20,
"score": 125886
}
] |
Bank distress in the news: Describing events through deep learning
|
[
{
"content": "Title: Learning Filter Banks Using Deep Learning For Acoustic Signals Content: Designing appropriate features for acoustic event recognition tasks is an active field of research. Expressive features should both improve the performance of the tasks and also be interpret-able. Currently, heuristically designed features based on the domain knowledge requires tremendous effort in hand-crafting, while features extracted through deep network are difficult for human to interpret. In this work, we explore the experience guided learning method for designing acoustic features. This is a novel hybrid approach combining both domain knowledge and purely data driven feature designing. Based on the procedure of log Mel-filter banks, we design a filter bank learning layer. We concatenate this layer with a convolutional neural network (CNN) model. After training the network, the weight of the filter bank learning layer is extracted to facilitate the design of acoustic features. We smooth the trained weight of the learning layer and re-initialize it in filter bank learning layer as audio feature extractor. For the environmental sound recognition task based on the Urbansound8K dataset [1], the experience guided learning leads to a 2% accuracy improvement compared with the fixed feature extractors (the log Mel-filter bank). The shape of the new filter banks are visualized and explained to prove the effectiveness of the feature design process.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "8f67ff7d7a4fc72d87f82ae340dba4365b7ea664",
"rank": 1,
"score": 148524
},
{
"content": "Title: Applying Deep Learning to the Newsvendor Problem Content: The newsvendor problem is one of the most basic and widely applied inventory models. There are numerous extensions of this problem. If the probability distribution of the demand is known, the problem can be solved analytically. However, approximating the probability distribution is not easy and is prone to error; therefore, the resulting solution to the newsvendor problem may be not optimal. To address this issue, we propose an algorithm based on deep learning that optimizes the order quantities for all products based on features of the demand data. Our algorithm integrates the forecasting and inventory-optimization steps, rather than solving them separately, as is typically done, and does not require knowledge of the probability distributions of the demand. Numerical experiments on real-world data suggest that our algorithm outperforms other approaches, including data-driven and machine learning approaches, especially for demands with high volatility. Finally, in order to show how this approach can be used for other inventory optimization problems, we provide an extension for (r,Q) policies.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "834a3bd7caa3fa967e124d04be64f038cf8951a7",
"rank": 2,
"score": 122021
},
{
"content": "Title: Deep Learning for Event-Driven Stock Prediction Content: Neural Tensor Network for Learning Event Embeddings Event Representation E = (O1, P, O2, T), where P is the action, O1 is the actor, O2 is the object and T is the timestamp (T is mainly used for aligning stock data with news data). For example, the event “Sep 3, 2013 Microsoft agrees to buy Nokia’s mobile phone business for $7.2 billion.” is modeled as: (Actor = Microsoft, Action = buy, Object = Nokia’s mobile phone business, Time = Sep 3, 2013) Event Embedding",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "4938e8c8c9ea3d351d283181819af5e5801efbed",
"rank": 3,
"score": 120587
},
{
"content": "Title: Unsupervised Feature Learning Based on Deep Models for Environmental Audio Tagging Content: Environmental audio tagging aims to predict only the presence or absence of certain acoustic events in the interested acoustic scene. In this paper, we make contributions to audio tagging in two parts, respectively, acoustic modeling and feature learning. We propose to use a shrinking deep neural network DNN framework incorporating unsupervised feature learning to handle the multilabel classification task. For the acoustic modeling, a large set of contextual frames of the chunk are fed into the DNN to perform a multilabel classification for the expected tags, considering that only chunk or utterance level rather than frame-level labels are available. Dropout and background noise aware training are also adopted to improve the generalization capability of the DNNs. For the unsupervised feature learning, we propose to use a symmetric or asymmetric deep denoising auto-encoder syDAE or asyDAE to generate new data-driven features from the logarithmic Mel-filter banks features. The new features, which are smoothed against background noise and more compact with contextual information, can further improve the performance of the DNN baseline. Compared with the standard Gaussian mixture model baseline of the DCASE 2016 audio tagging challenge, our proposed method obtains a significant equal error rate EER reduction from 0.21 to 0.13 on the development set. The proposed asyDAE system can get a relative 6.7% EER reduction compared with the strong DNN baseline on the development set. Finally, the results also show that our approach obtains the state-of-the-art performance with 0.15 EER on the evaluation set of the DCASE 2016 audio tagging task while EER of the first prize of this challenge is 0.17.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "f2d2dd3db244dcbc6fb32ff9c01ed0cdeb3fd437",
"rank": 4,
"score": 109537
},
{
"content": "Title: Listening to Chaotic Whispers: A Deep Learning Framework for News-oriented Stock Trend Prediction Content: Stock trend prediction plays a critical role in seeking maximized profit from the stock investment. However, precise trend prediction is very difficult since the highly volatile and non-stationary nature of the stock market. Exploding information on the Internet together with the advancing development of natural language processing and text mining techniques have enabled investors to unveil market trends and volatility from online content. Unfortunately, the quality, trustworthiness, and comprehensiveness of online content related to stock market vary drastically, and a large portion consists of the low-quality news, comments, or even rumors. To address this challenge, we imitate the learning process of human beings facing such chaotic online news, driven by three principles: sequential content dependency, diverse influence, and effective and efficient learning. In this paper, to capture the first two principles, we designed a Hybrid Attention Networks(HAN) to predict the stock trend based on the sequence of recent related news. Moreover, we apply the self-paced learning mechanism to imitate the third principle. Extensive experiments on real-world stock market data demonstrate the effectiveness of our framework. A further simulation illustrates that a straightforward trading strategy based on our proposed framework can significantly increase the annualized return.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "f2e9f869a9fc1f07887866be5f70a37b6c31411b",
"rank": 5,
"score": 107081
},
{
"content": "Title: Exploring Deep Learning Methods for Discovering Features in Speech Signals Content: Exploring Deep Learning Methods for discovering features in speech signals. Navdeep Jaitly Doctor of Philosophy Graduate Department of Computer Science University of Toronto 2014 This thesis makes three main contributions to the area of speech recognition with Deep Neural Network Hidden Markov Models (DNN-HMMs). Firstly, we explore the effectiveness of features learnt from speech databases using Deep Learning for speech recognition. This contrasts with prior works that have largely confined themselves to using traditional features such as Mel Cepstral Coefficients and Mel log filter banks for speech recognition. We start by showing that features learnt on raw signals using Gaussian-ReLU Restricted Boltzmann Machines can achieve accuracy close to that achieved with the best traditional features. These features are, however, learnt using a generative model that ignores domain knowledge. We develop methods to discover features that are endowed with meaningful semantics that are relevant to the domain using capsules. To this end, we extend previous work on transforming autoencoders and propose a new autoencoder with a domain-specific decoder to learn capsules from speech databases. We show that capsule instantiation parameters can be combined with Mel log filter banks to produce improvements in phone recognition on TIMIT. On WSJ the word error rate does not improve, even though we get strong gains in classification accuracy. We speculate this may be because of the mismatched objectives of word error rate over an utterance and frame error rate on the sub-phonetic class for a frame. Secondly, we develop a method for data augmentation in speech datasets. Such methods result in strong gains in object recognition, but have largely been ignored in speech recognition. Our data augmentation encourages the learning of invariance to vocal tract length of speakers. The method is shown to improve the phone error rate on TIMIT and the word error rate on a 14 hour subset of WSJ. Lastly, we develop a method for learning and using a longer range model of targets, conditioned on the input. This method predicts the labels for multiple frames together and uses a geometric average of these predictions during decoding. It produces state of the art results on phone recognition with TIMIT and also produces significant gains on WSJ.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "b04175bb99d6beff0f201ed82971aeb91d2c081d",
"rank": 6,
"score": 106778
},
{
"content": "Title: Deep Learning Content: Machine-learning technology powers many aspects of modern society: from web searches to content filtering on social networks to recommendations on e-commerce websites, and it is increasingly present in consumer products such as cameras and smartphones. Machine-learning systems are used to identify objects in images, transcribe speech into text, match news items, posts or products with users’ interests, and select relevant results of search. Increasingly, these applications make use of a class of techniques called deep learning. Conventional machine-learning techniques were limited in their ability to process natural data in their raw form. For decades, constructing a pattern-recognition or machine-learning system required careful engineering and considerable domain expertise to design a feature extractor that transformed the raw data (such as the pixel values of an image) into a suitable internal representation or feature vector from which the learning subsystem, often a classifier, could detect or classify patterns in the input. Representation learning is a set of methods that allows a machine to be fed with raw data and to automatically discover the representations needed for detection or classification. Deep-learning methods are representation-learning methods with multiple levels of representation, obtained by composing simple but non-linear modules that each transform the representation at one level (starting with the raw input) into a representation at a higher, slightly more abstract level. With the composition of enough such transformations, very complex functions can be learned. For classification tasks, higher layers of representation amplify aspects of the input that are important for discrimination and suppress irrelevant variations. An image, for example, comes in the form of an array of pixel values, and the learned features in the first layer of representation typically represent the presence or absence of edges at particular orientations and locations in the image. The second layer typically detects motifs by spotting particular arrangements of edges, regardless of small variations in the edge positions. The third layer may assemble motifs into larger combinations that correspond to parts of familiar objects, and subsequent layers would detect objects as combinations of these parts. The key aspect of deep learning is that these layers of features are not designed by human engineers: they are learned from data using a general-purpose learning procedure. Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence community for many years. It has turned out to be very good at discovering intricate structures in high-dimensional data and is therefore applicable to many domains of science, business and government. In addition to beating records in image recognition and speech recognition, it has beaten other machine-learning techniques at predicting the activity of potential drug molecules, analysing particle accelerator data, reconstructing brain circuits, and predicting the effects of mutations in non-coding DNA on gene expression and disease. Perhaps more surprisingly, deep learning has produced extremely promising results for various tasks in natural language understanding, particularly topic classification, sentiment analysis, question answering and language translation. We think that deep learning will have many more successes in the near future because it requires very little engineering by hand, so it can easily take advantage of increases in the amount of available computation and data. New learning algorithms and architectures that are currently being developed for deep neural networks will only accelerate this progress.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "15cf63f8d44179423b4100531db4bb84245aa6f1",
"rank": 7,
"score": 105546
},
{
"content": "Title: Deep Learning for Imbalanced Multimedia Data Classification Content: Classification of imbalanced data is an important research problem as lots of real-world data sets have skewed class distributions in which the majority of data instances (examples) belong to one class and far fewer instances belong to others. While in many applications, the minority instances actually represent the concept of interest (e.g., fraud in banking operations, abnormal cell in medical data, etc.), a classifier induced from an imbalanced data set is more likely to be biased towards the majority class and show very poor classification accuracy on the minority class. Despite extensive research efforts, imbalanced data classification remains one of the most challenging problems in data mining and machine learning, especially for multimedia data. To tackle this challenge, in this paper, we propose an extended deep learning approach to achieve promising performance in classifying skewed multimedia data sets. Specifically, we investigate the integration of bootstrapping methods and a state-of-the-art deep learning approach, Convolutional Neural Networks (CNNs), with extensive empirical studies. Considering the fact that deep learning approaches such as CNNs are usually computationally expensive, we propose to feed low-level features to CNNs and prove its feasibility in achieving promising performance while saving a lot of training time. The experimental results show the effectiveness of our framework in classifying severely imbalanced data in the TRECVID data set.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "42ef50955a7f12afad78f0bd3819dbc555580225",
"rank": 8,
"score": 104351
},
{
"content": "Title: Using deep learning for short text understanding Content: Classifying short texts to one category or clustering semantically related texts is challenging, and the importance of both is growing due to the rise of microblogging platforms, digital news feeds, and the like. We can accomplish this classifying and clustering with the help of a deep neural network which produces compact binary representations of a short text, and can assign the same category to texts that have similar binary representations. But problems arise when there is little contextual information on the short texts, which makes it difficult for the deep neural network to produce similar binary codes for semantically related texts. We propose to address this issue using semantic enrichment. This is accomplished by taking the nouns, and verbs used in the short texts and generating the concepts and co-occurring words with the help of those terms. The nouns are used to generate concepts within the given short text, whereas the verbs are used to prune the ambiguous context (if any) present in the text. The enriched text then goes through a deep neural network to produce a prediction label for that short text representing it’s category.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "2b337d6a72c8c2b1d97097dc24ec0e9a8d4c2186",
"rank": 9,
"score": 101490
},
{
"content": "Title: PCANet: A Simple Deep Learning Baseline for Image Classification? Content: In this paper, we propose a very simple deep learning network for image classification that is based on very basic data processing components: 1) cascaded principal component analysis (PCA); 2) binary hashing; and 3) blockwise histograms. In the proposed architecture, the PCA is employed to learn multistage filter banks. This is followed by simple binary hashing and block histograms for indexing and pooling. This architecture is thus called the PCA network (PCANet) and can be extremely easily and efficiently designed and learned. For comparison and to provide a better understanding, we also introduce and study two simple variations of PCANet: 1) RandNet and 2) LDANet. They share the same topology as PCANet, but their cascaded filters are either randomly selected or learned from linear discriminant analysis. We have extensively tested these basic networks on many benchmark visual data sets for different tasks, including Labeled Faces in the Wild (LFW) for face verification; the MultiPIE, Extended Yale B, AR, Facial Recognition Technology (FERET) data sets for face recognition; and MNIST for hand-written digit recognition. Surprisingly, for all tasks, such a seemingly naive PCANet model is on par with the state-of-the-art features either prefixed, highly hand-crafted, or carefully learned [by deep neural networks (DNNs)]. Even more surprisingly, the model sets new records for many classification tasks on the Extended Yale B, AR, and FERET data sets and on MNIST variations. Additional experiments on other public data sets also demonstrate the potential of PCANet to serve as a simple but highly competitive baseline for texture classification and object recognition.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "841a5de1d71a0b51957d9be9d9bebed33fb5d9fa",
"rank": 10,
"score": 101380
},
{
"content": "Title: Fake News Detection on Social Media using Geometric Deep Learning Content: Social media are nowadays one of the main news sources for millions of people around the globe due to their low cost, easy access, and rapid dissemination. This however comes at the cost of dubious trustworthiness and significant risk of exposure to ‘fake news’, intentionally written to mislead the readers. Automatically detecting fake news poses challenges that defy existing content-based analysis approaches. One of the main reasons is that often the interpretation of the news requires the knowledge of political or social context or ‘common sense’, which current natural language processing algorithms are still missing. Recent studies have empirically shown that fake and real news spread differently on social media, forming propagation patterns that could be harnessed for the automatic fake news detection. Propagation-based approaches have multiple advantages compared to their content-based counterparts, among which is language independence and better resilience to adversarial attacks. In this paper, we show a novel automatic fake news detection model based on geometric deep learning. The underlying core algorithms are a generalization of classical convolutional neural networks to graphs, allowing the fusion of heterogeneous data such as content, user profile and activity, social graph, and news propagation. Our model was trained and tested on news stories, verified by professional fact-checking organizations, that were spread on Twitter. Our experiments indicate that social network structure and propagation are important features allowing highly accurate (92.7% ROC AUC) fake news detection. Second, we observe that fake news can be reliably detected at an early stage, after just a few hours of propagation. Third, we test the aging of our model on training and testing data separated in time. Our results point to the promise of propagation-based approaches for fake news detection as an alternative or complementary strategy to content-based approaches. ar X iv :1 90 2. 06 67 3v 1 [ cs .S I] 1 0 Fe b 20 19",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "318d7da35307221267b6ce6ead995cc812245abb",
"rank": 11,
"score": 99900
},
{
"content": "Title: Deep learning with support vector data description Content: One of the most critical problems for machine learning methods is overfitting. The overfitting problem is a phenomenon in which the accuracy of the model on unseen data is poor whereas the training accuracy is nearly perfect. This problem is particularly severe in complex models that have a large set of parameters. In this paper, we propose a deep learning neural network model that adopts the support vector data description (SVDD). The SVDD is a variant of the support vector machine, which has high generalization performance by acquiring a maximal margin in one-class classification problems. The proposed model strives to obtain the representational power of deep learning. Generalization performance is maintained using the SVDD. The experimental results showed that the proposed model can learn multiclass data without severe overfitting problems.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "1ba1de0f143bd3166c9961acc869e123651d9836",
"rank": 12,
"score": 99008
},
{
"content": "Title: Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique Content: D EEP learning is a growing trend in general data analysis and has been termed one of the 10 breakthrough technologies of 2013 [1]. Deep learning is an improvement of artificial neural networks, consisting of more layers that permit higher levels of abstraction and improved predictions from data [2]. To date, it is emerging as the leading machine-learning tool in the general imaging and computer vision domains. In particular, convolutional neural networks (CNNs) have proven to be powerful tools for a broad range of computer vision tasks. Deep CNNs automatically learn mid-level and high-level abstractions obtained from raw data (e.g., images). Recent results indicate that the generic descriptors extracted from CNNs are extremely effective in object recognition and localization in natural images. Medical image analysis groups across the world are quickly entering the field and applying CNNs and other deep learning methodologies to a wide variety of applications. Promising results are emerging. In medical imaging, the accurate diagnosis and/or assessment of a disease depends on both image acquisition and image interpretation. Image acquisition has improved substantially over recent years, with devices acquiring data at faster rates and increased resolution. The image interpretation process, however, has only recently begun to benefit from computer technology. Most interpretations of medical images are performed by physicians; however, image interpretation by humans is limited due to its subjectivity, large variations across interpreters, and fatigue. Many diagnostic tasks require an initial search process to detect abnormalities, and to quantify measurements and changes over time. Computerized tools, specifically image analysis and machine learning, are the key enablers to improve diagnosis, by facilitating identification of the findings that require treatment and to support the expert’s workflow. Among these tools, deep learning is rapidly proving to be the state-of-the-art foundation, leading to improved accuracy. It has also opened up new frontiers in data analysis with rates of progress not before experienced.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "0ab99aa04e3a8340a7552355fb547374a5604b24",
"rank": 13,
"score": 97883
},
{
"content": "Title: New York University 2016 System for KBP Event Nugget: A Deep Learning Approach Content: This is the first time New York University (NYU) participates in the event nugget (EN) evaluation of the Text Analysis Conference (TAC). We developed EN systems for both subtasks of event nugget, i.e, EN Task 1: Event Nugget Detection and EN Task 2: Event Nugget Detection and Coreference. The systems are mainly based on our recent research on deep learning for event detection (Nguyen and Grishman, 2015a; Nguyen and Grishman, 2016a). Due to the limited time we could devote to system development this year, we only ran the systems on the English evaluation data. However, we expect that the adaptation of the current systems to new languages can be done quickly. The development experiments show that although our current systems do not rely on complicated feature engineering, they significantly outperform the reported systems last year for the EN subtasks on the 2015 evaluation data.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "a6df9a75a7a946cad8c32ee2a8c88d826a21430c",
"rank": 14,
"score": 96364
},
{
"content": "Title: Clothing identification via deep learning: forensic applications Content: Attribute-based identification systems are essential for forensic investigations because they help in identifying individuals. An item such as clothing is a visual attribute because it can usually be used to describe people. The method proposed in this article aims to identify people based on the visual information derived from their attire. Deep learning is used to train the computer to classify images based on clothing content. We first demonstrate clothing classification using a large scale dataset, where the proposed model performs relatively poorly. Then, we use clothing classification on a dataset containing popular logos and famous brand images. The results show that the model correctly classifies most of the test images with a success rate that is higher than 70%. Finally, we evaluate clothing classification using footage from surveillance cameras. The system performs well on this dataset, labelling 70% of the test images correctly.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "b53a55d624269228a761301abab1d80ade9f89c5",
"rank": 15,
"score": 95960
},
{
"content": "Title: Deep Machine Learning and Neural Networks: An Overview Content: Received Feb 10, 2017 Revised Apr 14, 2017 Accepted May 23, 2017 Deep learning is a technique of machine learning in artificial intelligence area. Deep learning in a refined \"machine learning\" algorithm that far surpasses a considerable lot of its forerunners in its capacities to perceive syllables and picture. Deep learning is as of now a greatly dynamic examination territory in machine learning and example acknowledgment society. It has increased colossal triumphs in an expansive zone of utilizations, for example, speech recognition, computer vision and natural language processing and numerous industry item. Neural network is used to implement the machine learning or to design intelligent machines. In this paper brief introduction to all machine learning paradigm and application area of deep machine learning and different types of neural networks with applications is discussed. Keyword:",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "d101040124ccceb0cd46f92c8815d5c475605cd1",
"rank": 16,
"score": 95534
},
{
"content": "Title: 3D Deep Learning for Efficient and Robust Landmark Detection in Volumetric Data Content: Recently, deep learning has demonstrated great success in computer vision with the capability to learn powerful image features from a large training set. However, most of the published work has been confined to solving 2D problems, with a few limited exceptions that treated the 3D space as a composition of 2D orthogonal planes. The challenge of 3D deep learning is due to a much larger input vector, compared to 2D, which dramatically increases the computation time and the chance of over-fitting, especially when combined with limited training samples (hundreds to thousands), typical for medical imaging applications. To address this challenge, we propose an efficient and robust deep learning algorithm capable of full 3D detection in volumetric data. A two-step approach is exploited for efficient detection. A shallow network (with one hidden layer) is used for the initial testing of all voxels to obtain a small number of promising candidates, followed by more accurate classification with a deep network. In addition, we propose two approaches, i.e., separable filter decomposition and network sparsification, to speed up the evaluation of a network. To mitigate the over-fitting issue, thereby increasing detection robustness, we extract small 3D patches from a multi-resolution image pyramid. The deeply learned image features are further combined with Haar wavelet features to increase the detection accuracy. The proposed method has been quantitatively evaluated for carotid artery bifurcation detection on a head-neck CT dataset from 455 patients. Compared to the state-ofthe-art, the mean error is reduced by more than half, from 5.97 mm to 2.64 mm, with a detection speed of less than 1 s/volume.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "6cd3ea4361e035969e6cf819422d0262f7c0a186",
"rank": 17,
"score": 95451
},
{
"content": "Title: Deep Learning for Wind Speed Forecasting in Northeastern Region of Brazil Content: Deep Learning is one of the latest approaches in the field of artificial neural networks. Since they were first proposed in mid-2006, Deep Learning models have obtained state-of-art results in some problems with classification and pattern recognition. However, such models have been little used in time series forecasting. This work aims to investigate the use of some of these architectures in this kind of problem, specifically in predicting the hourly average speed of winds in the Northeastern region of Brazil. The results showed that Deep Learning offers a good alternative for performing this task, overcoming some results of previous works.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "87094ebab924f893160716021a8a5bc645b3ff1f",
"rank": 18,
"score": 95332
},
{
"content": "Title: Deep Learning in Radiology: Does One Size Fit All? Content: Deep learning (DL) is a popular method that is used to perform many important tasks in radiology and medical imaging. Some forms of DL are able to accurately segment organs (essentially, trace the boundaries, enabling volume measurements or calculation of other properties). Other DL networks are able to predict important properties from regions of an image-for instance, whether something is malignant, molecular markers for tissue in a region, even prognostic markers. DL is easier to train than traditional machine learning methods, but requires more data and much more care in analyzing results. It will automatically find the features of importance, but understanding what those features are can be a challenge. This article describes the basic concepts of DL systems and some of the traps that exist in building DL systems and how to identify those traps.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "2301f3bebd0cebbf161a017bbb70faffbb2c2506",
"rank": 19,
"score": 95123
},
{
"content": "Title: Transductive Event Classification through Heterogeneous Networks Content: Events can be defined as \"something that occurs at specific place and time associated with some specific actions\". In general, events extracted from news articles and social networks are used to map the information from web to the various phenomena that occur in our physical world. One of the main steps to perform this relationship is the use of machine learning algorithms for event classification, which has received great attention in the web document engineering field in recent years. Traditional machine learning algorithms are based on vector space model representations and supervised classification. However, events are composed of multiple representations such as textual data, temporal information, geographic location and other types of metadata. All these representations are poorly represented together in a vector space model. Moreover, supervised classification requires the labeling of a significant sample of events to construct a training set for learning process, thereby hampering the practical application of event classification. In this paper, we propose a method called TECHN (Transductive Event Classification through Heterogeneous Networks), which considers event metadata as different objects in an heterogeneous network. Besides, the TECHN method has the ability to automatically learn which types of network objects (event metadata) are most efficient in the classification task. In addition, our TECHN method is based on a transductive classification that considers both labeled events and a vast amount of unlabeled events. The experimental results show that TECHN method obtains promising results, especially when we consider different weights of importance for each type of event metadata and a small set of labeled events.",
"qid": "01d208b33561362f7714f714d3bc4a1f7aa1637c",
"docid": "54598a8872516bf85b9dbe79bfa573351b0f2042",
"rank": 20,
"score": 94338
}
] |
Future Perspectives on Next Generation e-Sports Infrastructure and Exploring Their Benefits
|
[
{
"content": "Title: Analysing B2B electronic procurement benefits: information systems perspective Content: Abstract Purpose – This paper aims to present electronic procurement benefits identified in four case companies from the information technology (IT), hi-tech sector. Design/methodology/approach – Multi-case study design was applied. The benefits reported in the companies were analysed and classified according to taxonomies from the information systems discipline. Finally, a new benefits classification was proposed. The framework was developed based on information systems literature. Findings – The research confirmed difficulties with benefits evaluation, as, apart from operational benefits, non-financial, intangible benefits at strategic level were also identified. Traditional evaluation methods are unable to capture all benefits categories, especially at strategic level. New taxonomy was created, which allows evaluation of the complex e-procurement impact. In the proposed taxonomy, e-procurement benefits are classified according to their level (operational, tactical, strategic), area of impact, applying scorecard dimensions (customer, process, financial, learning and growth). In addition the benefits characteristic is captured (tangible, intangible, financial and non-financial). Research limitations/implications – Research is based on four case studies only. Findings are specific to case companies and the environment in which they operate. The framework should be tested further in different contexts. Practical implications – The new taxonomy allows evaluation of the complex e-procurement impact, demonstrating that benefits achieved do not concern merely the financial impact. The framework can be applied to preparing new systems implementation as well as to evaluating existing systems. Originality/value – The paper applies information systems frameworks to the electronic procurement field, which allows one to look at e-procurement systems considering its complex impact. The framework can also be used to evaluate different systems, not simply e-procurement.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "b3e3f63c25e22d5a13cb3d73561fc752b913b7dd",
"rank": 1,
"score": 91695
},
{
"content": "Title: Infrastructure support for mobile computing Content: Mobile computing is emerging as the prime focus of next generation computing .One of the prime issues of mobile computing is to provide infrastructure support in terms of computing devices, seamless mobility, application middleware, data and user security, and user applications/services. Mobile commerce is one of the driving forces that has evinced enormous interest in mobile computing .The thought of conducting commerce on the go is what is driving the huge investments corporations are making in researching this area. This paper discusses the various challenges in providing infrastructure for wireless computing.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "047eeb7fce304fdb2b41f3c4d0b393dd1137bdab",
"rank": 2,
"score": 90819
},
{
"content": "Title: Service-oriented paradigms in industrial automation Content: This paper outlines opportunities and challenges in the development of next-generation embedded devices, applications, and services, resulting from their increasing intelligence - it plots envisioned future directions for intelligent device networking based on service-oriented high-level protocols, in particular as regards the industrial automation sector - and outlines the approach adopted by the Service Infrastructure for Real-Time Embedded Networked Applications project, as well as the business advantages this approach is expected to provide.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "1474f7f3d8c13b5fa0f725b55e6baf736b9e4bfc",
"rank": 3,
"score": 86442
},
{
"content": "Title: New Generation Sensor Web Enablement Content: Many sensor networks have been deployed to monitor Earth's environment, and more will follow in the future. Environmental sensors have improved continuously by becoming smaller, cheaper, and more intelligent. Due to the large number of sensor manufacturers and differing accompanying protocols, integrating diverse sensors into observation systems is not straightforward. A coherent infrastructure is needed to treat sensors in an interoperable, platform-independent and uniform way. The concept of the Sensor Web reflects such a kind of infrastructure for sharing, finding, and accessing sensors and their data across different applications. It hides the heterogeneous sensor hardware and communication protocols from the applications built on top of it. The Sensor Web Enablement initiative of the Open Geospatial Consortium standardizes web service interfaces and data encodings which can be used as building blocks for a Sensor Web. This article illustrates and analyzes the recent developments of the new generation of the Sensor Web Enablement specification framework. Further, we relate the Sensor Web to other emerging concepts such as the Web of Things and point out challenges and resulting future work topics for research on Sensor Web Enablement.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "961b8e95e4b360e5d95ef79a21958540d4e551ab",
"rank": 4,
"score": 84959
},
{
"content": "Title: Rethink fronthaul for soft RAN Content: In this article we discuss the design of a new fronthaul interface for future 5G networks. The major shortcomings of current fronthaul solutions are first analyzed, and then a new fronthaul interface called next-generation fronthaul interface (NGFI) is proposed. The design principles for NGFI are presented, including decoupling the fronthaul bandwidth from the number of antennas, decoupling cell and user equipment processing, and focusing on high-performancegain collaborative technologies. NGFI aims to better support key 5G technologies, in particular cloud RAN, network functions virtualization, and large-scale antenna systems. NGFI claims the advantages of reduced bandwidth as well as improved transmission efficiency by exploiting the tidal wave effect on mobile network traffic. The transmission of NGFI is based on Ethernet to enjoy the benefits of flexibility and reliability. The major impact, challenges, and potential solutions of Ethernet-based fronthaul networks are also analyzed. Jitter, latency, and time and frequency synchronization are the major issues to overcome.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "45654695f5cad20d2be36d45d280af5180004baf",
"rank": 5,
"score": 81758
},
{
"content": "Title: The Complementarity of Information Technology Infrastructure and E-Commerce Capability: A Resource-Based Assessment of Their Business Value Content: This study seeks to assess the business value of e-commerce capability and information technology (IT) infrastructure in the context of electronic business at the firm level. Grounded in the IT business-value literature and enhanced by the resource-based theory of the firm, we developed a research framework in which both the main effects and the interaction effects of e-commerce and IT on firm performance were tested. Within this theoretical framework, we formulated several hypotheses. We then developed a multidimensional e-commerce capability construct, and after establishing its validity and reliability, tested the hypotheses with empirical data from 114 companies in the retail industry. Controlling for variations of firm size and subindustry effects, our empirical analysis found a strong positive interaction effect between IT infrastructure and e-commerce capability. This suggests that their complementarity positively contributes to firm performance in terms of sales per employee, inventory turnover, and cost reduction. The results are consistent with the resource-based theory, and provide empirical evidence to the complementary synergy between front-end e-commerce capability and back-end IT infrastructure. Combined together, they become more effective in producing business value. Yet the value of this synergy has not been recognized in the IT payoff literature. The “productivity paradox” observed in various studies has been attributed to variation in methods and measures, yet we offer an additional explanation: ignoring complementarities in business value measurement implies that the impact of IT was seriously underestimated. Our results emphasized the integration of resources as a feasible path to e-commerce value—companies need to enhance the integration between front-end e-commerce capability and back-end IT infrastructure in order to reap the benefits of e-commerce investments.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "10e887664b0e9d0b193e8dc8e92f6f1a8bfba5e5",
"rank": 6,
"score": 81737
},
{
"content": "Title: An information-centric energy infrastructure: The Berkeley view Content: We describe an approach for how to design an essentially more scalable, flexible and resilient electric power infrastructure – one that encourages efficient use, integrates local generation, and manages demand through omnipresent awareness of energy availability and use over time. We are inspired by how the Internet has revolutionized communications infrastructure, by pushing intelligence to the edges while hiding the diversity of underlying technologies through well-defined interfaces. Any end device is a traffic source or sink and intelligent endpoints adapt their traffic to what the infrastructure can",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "7b9ab27ad78899b6b284a17c38aa75fb0e1d1765",
"rank": 7,
"score": 80161
},
{
"content": "Title: Performance Evaluation and Optimization of Communication Infrastructure for the Next Generation Air Transportation System Content: Automatic dependent surveillance-broadcast (ADS-B) is one of the fundamental surveillance technologies to improve the safety, capacity, and efficiency of the national airspace system. ADS-B shares its frequency band with current radar systems that use the same 1,090 MHz band. The coexistence of radar systems and ADS-B systems is a key issue to detect and resolve conflicts in the next generation air transportation system (NextGen). This paper focuses on the performance evaluation of ADS-B with existing radar systems and performance optimization of ADS-B systems to improve the safety and efficiency of conflict detection and resolution in NextGen. We have developed a simulation environment which models the complex interplay among the air traffic load, the radar systems, the ADS-B systems, and the wireless channel. A simple model is used to derive an analytical expression for a performance metric of ADS-B. This model is then used to design an adaptive ADS-B protocol for maximizing the information coverage while guaranteeing reliable and timely communication in air traffic surveillance networks. Simulation results show that the effect of ADS-B interference on the current radar system is negligible. The operational ability of ADS-B meets the performance requirements of conflict detection and resolution in air traffic control. However, upgrades are required in the current radar system for operation within an ADS-B environment since the current radars can significantly degrade the ADS-B performance. Numerical results indicate that the proposed adaptive protocol has the potential to improve the performance of conflict detection and resolution in air traffic control.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "0fe3a88a22a018651e4886d547488e5a3ce224fc",
"rank": 8,
"score": 79623
},
{
"content": "Title: Architecting the next generation of service-based SCADA/DCS system of systems Content: SCADA and DCS systems are in the heart of the modern industrial infrastructure. The rapid changes in the networked embedded systems and the way industrial applications are designed and implemented, call for a shift in the architectural paradigm. Next generation SCADA and DCS systems will be able to foster cross-layer collaboration with the shop-floor devices as well as in-network and enterprise applications. Ecosystems driven by (web) service based interactions will enable stronger coupling of real-world and the business side, leading to a new generation of monitoring and control applications and services witnessed as the integration of large-scale systems of systems that are constantly evolving to address new user needs.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "875e08da83c0d499da9d9a5728d492d35d96773c",
"rank": 9,
"score": 79177
},
{
"content": "Title: E-government implementation: A bird's eye view of issues relating to costs, opportunities, benefits and risks Content: After more than a decade of comprehensive research work in the area of electronic government (e-government), no attempt has yet been made to undertake a systematic literature review on the costs, opportunities, benefits and risks that influence the implementation of e-government. This is particularly significant given the various related challenges that governments have faced over the years when implementing e-government initiatives. Hence, the aim of this paper is to undertake a comprehensive analysis of relevant literature addressing these issues using a systematic review of 132 studies identified from the Scopus online database and Google Scholar together with a manual review of relevant papers from journals dedicated to electronic government research such as Electronic Government, an International Journal (EGIJ), International Journal of Electronic Government Research (IJEGR) and Transforming Government: People, Process, and Policy (TGPPP). The overall review indicated that although a large number of papers discuss costs, opportunities, benefits and risks, treatment of these issues have tended to be superficial.Moreover, there is a lack of empirical studies which can statistically evaluate the performance of these constructs in relation to the various egovernment systems. Therefore, this research would help governments to better analyse the impact of costs, opportunities, benefits and risks on the success of e-government systems and its pre-adoption from an implementation perspective.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "4928aee4b9a558d8faaa6126201a45b7aaea7bb6",
"rank": 10,
"score": 77962
},
{
"content": "Title: Cyber security in the Smart Grid: Survey and challenges Content: The Smart Grid, generally referred to as the next-generation power system, is considered as a revolutionary and evolutionary regime of existing power grids. More importantly, with the integration of advanced computing and communication technologies, the Smart Grid is expected to greatly enhance efficiency and reliability of future power systems with renewable energy resources, as well as distributed intelligence and demand response. Along with the silent features of the Smart Grid, cyber security emerges to be a critical issue because millions of electronic devices are inter-connected via communication networks throughout critical power facilities, which has an immediate impact on reliability of such a widespread infrastructure. In this paper, we present a comprehensive survey of cyber security issues for the Smart Grid. Specifically, we focus on reviewing and discussing security requirements, network vulnerabilities, attack countermeasures, secure communication protocols and architectures in the Smart Grid. We aim to provide a deep understanding of security vulnerabilities and solutions in the Smart Grid and shed light on future research directions for Smart Grid security. 2013 Elsevier B.V. All rights reserved.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "0b458ce6c0d6d7fd20499e5b64a46132d7c380f2",
"rank": 11,
"score": 77669
},
{
"content": "Title: Neural Networks and Graph Algorithms with Next-Generation Processors Content: The use of graphical processors for distributed computation revolutionized the field of high performance scientific computing. As the Moore's Law era of computing draws to a close, the development of non-Von Neumann systems: neuromorphic processing units, and quantum annealers; again are redefining new territory for computational methods. While these technologies are still in their nascent stages, we discuss their potential to advance computing in two domains: machine learning, and solving constraint satisfaction problems. Each of these processors utilize fundamentally different theoretical models of computation. This raises questions about how to best use them in the design and implementation of applications. While many processors are being developed with a specific domain target, the ubiquity of spin-glass models and neural networks provides an avenue for multi-functional applications. This provides hints at the future infrastructure needed to integrate many next-generation processing units into conventional high-performance computing systems.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "724e3c4e98bc9ac3281d5aef7d53ecfd4233c3fc",
"rank": 12,
"score": 77165
},
{
"content": "Title: Constraints in the IoT: The World in 2020 and Beyond Content: The Internet of Things (IoT), often referred as the future Internet; is a collection of interconnected devices integrated into the world-wide network that covers almost everything and could be available anywhere. IoT is an emerging technology and aims to play an important role in saving money, conserving energy, eliminating gap and better monitoring for intensive management on a routine basis. On the other hand, it is also facing certain design constraints such as technical challenges, social challenges, compromising privacy and performance tradeoffs. This paper surveys major technical limitations that are hindering the successful deployment of the IoT such as standardization, interoperability, networking issues, addressing and sensing issues, power and storage restrictions, privacy and security, etc. This paper categorizes the existing research on the technical constraints that have been published in the recent years. With this categorization, we aim to provide an easy and concise view of the technical aspects of the IoT. Furthermore, we forecast the changes influenced by the IoT. This paper predicts the future and provides an estimation of the world in year 2020 and beyond. Keywords—Internet of Things; Future Internet; Next generation network issues; World-wide network; 2020",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "5f2a4982a8adef2d1a6d589a155143291d440c0a",
"rank": 13,
"score": 75460
},
{
"content": "Title: A Resource-Based Model For E-Commerce In Developing Countries Content: Previous efforts in electronic commerce (e-commerce) research in developing countries shows that there is an acute lack of theoretical frameworks and empirical evidence to understand how developing country firms realize electronic commerce benefits amidst their national constraints. This paper sets out to develop a theoretically abstracted but contextually grounded model of how developing country firms can orient their resources and realize these benefits amidst their national constraints. A review of e-commerce and strategy management literature to develop a resource – based model for e-commerce benefits was undertaken. The process-based model provides an understanding of how to identify, integrate, and reconfigure resources to achieve electronic commerce benefits; provides propositions that serves as theoretical platforms for future empirically grounded research on electronic commerce in developing country contexts and brings organizations closer to identifying and categorizing the strategic value of resources and the role managerial capabilities and intangible resources play in sustaining e-commerce benefits. Finally, our findings provides organizations the strategic options to address resources which have lost their value or have become less valuable to their strategic orientation in e-commerce adoption thereby serving as a starting point of examining e-commerce in developing countries through the theoretical lens of information systems and strategic management.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "f1e0a619b6ad652b65b49f362ac9413e89291ad7",
"rank": 14,
"score": 73699
},
{
"content": "Title: Rocky 7: a next generation Mars rover prototype Content: This paper provides a system overview of a new Mars rover prototype, Rocky 7 1. We describe all system aspects: mechanical and electrical design, computer and software infrastructure , algorithms for navigation and manipulation, science data acquisition, and outdoor rover testing. In each area, the improved or added functionality is explained in a context of its path to ight, and within the constraints of desired science missions.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "4cd3cb09bb9d30c9cb5a74f4257a49ebe798ab79",
"rank": 15,
"score": 73689
},
{
"content": "Title: An image encryption scheme based on elliptic curve pseudo random and Advanced Encryption System Content: Elliptic curve cryptography (ECC) has proven to be an effective cryptography. ECC has its own advantages such as efficient key size compared to other public key infrastructures. This paper exploits the Elliptic curve random generator defined by National Institute of Standards and Technology (NIST) to generate a sequence of arbitrary numbers based on curves. The random generation phase is based on public shared key and a changing point G, which is a generator of a curve to obtain random sequences. Then, Advanced Encryption System is applied to these sequences acquiring arbitrary keys for encrypting image. Using AES alongside well distributed randoms provides a prominent encryption technique. Our experiments show that the proposed method fulfills the basics of cryptography including simpleness and correctness. Moreover, the results of the evaluation prove the effectiveness and security of the proposed method.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "d8d160d9f6e987aa9e95a2480e97105fdeebca8f",
"rank": 16,
"score": 73665
},
{
"content": "Title: An Efficient and Secure Automotive Wireless Software Update Framework Content: Future vehicles will be wirelessly connected to nearby vehicles, to the road infrastructure, and to the Internet, thereby becoming an integral part of the Internet of Things. New comfort features, safety functions, and a number of new vehicle-specific services will be integrated in future smart vehicles. These include a fast, secure, and reliable way to diagnose and reconfigure a vehicle, as well as the installation of new software (SW) on its integrated electronic control units (ECUs). Such wireless SW updates are beneficial for both automotive carmakers and customers, as they allow us to securely enable new features on the vehicle and to fix SW bugs by installing a new SW version over the air. A secure and dependable wireless SW update process is valuable in the entire lifetime of a modern vehicle as it can be used already during vehicle development and manufacturing process on the assembly line, as well as during vehicle maintenance in a service center. Additionally, future vehicles will allow us to remotely download up-to-date SW on the ECUs. To support this process over the entire vehicle's lifetime, a generic framework is needed. In this paper, SecUp, a generic framework enabling secure and efficient wireless automotive SW updates is proposed. SecUp utilizes IEEE 802.11s as wireless medium to interconnect vehicles and diagnostic devices in a dependable and fast way. Additionally, SecUp is enabling beneficial wireless SW update features such as parallel and partial SW updates to increase the efficiency, and comprises advanced security mechanisms to prevent abuse and attacks.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "e25e0ccdb347f85dda2725c1d37fc1790111c6af",
"rank": 17,
"score": 73658
},
{
"content": "Title: E-GOVERNMENT EVALUATION FACTORS: CITIZEN’S PERSPECTIVE Content: The e-government field is growing to a considerable size, both in its contents and position with respect to other research fields. The government to citizen segment of egovernment is taking the lead in terms of its importance and size. Like the evaluation of all other information systems initiatives, the evaluation of egovernments in both theory and practice has proved to be important but complex. The complexity of evaluation is mostly due to the multiple perspectives involved, the difficulties of quantifying benefits, and the social and technical context of use. The importance of e-government evaluation is due to the enormous investment of governments on delivering e-government services, and to the considerable pace of growing in the e-government field. However, despite the importance of the evaluation of e-government services, literature shows that e-government evaluation is still an immature area in terms of development and management. This work is part of a research effort that aims to develop a holistic evaluation framework for e-government systems. The main aim of this paper is to investigate the citizen’ perspective in evaluating e-government services, and present a set of evaluating factors that influence citizens’ utilization of e-government services. These evaluation factors can serve as part of an e-government evaluation framework. Moreover, the evaluation factors can also be used as means of providing valuable feedback for the planning of future egovernment initiatives.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "ee24da4eadc2fb1bd4fa881de108ff8ede664ff5",
"rank": 18,
"score": 73273
},
{
"content": "Title: A Survey of Mobile Cloud Computing Applications: Perspectives and Challenges Content: As mobile computing has been developed for decades, a new model for mobile computing, namely, mobile cloud computing, emerges resulting from the marriage of powerful yet affordable mobile devices and cloud computing. In this paper we survey existing mobile cloud computing applications, as well as speculate future generation mobile cloud computing applications. We provide insights for the enabling technologies and challenges that lie ahead for us to move forward from mobile computing to mobile cloud computing for building the next generation mobile cloud applications. For each of the challenges, we provide a survey of existing solutions, identify research gaps, and suggest future research areas.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "d247b14fef93a2d6e87a555b0b992a41996a387f",
"rank": 19,
"score": 73098
},
{
"content": "Title: RF modules (Tx-Rx) with multifunctional MMICs Content: Next generation RF sensor modules for multifunction active electronically steered antenna (AESA) systems will need a combination of different operating modes, such as radar, electronic warfare (EW) functionalities and communications/datalinks within the same antenna frontend. They typically operate in C-Band, X-Band and Ku-Band and imply a bandwidth requirement of more than 10 GHz. For the realisation of modern active electronically steered antennas, the transmit/receive (T/R) modules have to match strict geometry demands. A major challenge for these future multifunction RF sensor modules is dictated by the half-wavelength antenna grid spacing, that limits the physical channel width to < 12 mm or even less, depending on the highest frequency of operation with accordant beam pointing requirements. A promising solution to overcome these geometry demands is the reduction of the total monolithic microwave integrated circuit (MMIC) chip area, achieved by integrating individual RF functionalities, which are commonly achieved through individual integrated circuits (ICs), into new multifunctional (MFC) MMICs. Various concepts, some of them already implemented, towards next generation RF sensor modules will be discussed and explained in this work.",
"qid": "0210b3fe6f7173c86936b5dd9261bc0be0c45652",
"docid": "9e5158222c911bec96d4f533cd0d7a1a0cff1731",
"rank": 20,
"score": 72295
}
] |
Pooled motion features for first-person videos
|
[
{
"content": "Title: Action and Interaction Recognition in First-Person Videos Content: In this work, we evaluate the performance of the popular dense trajectories approach on first-person action recognition datasets. A person moving around with a wearable camera will actively interact with humans and objects and also passively observe others interacting. Hence, in order to represent real-world scenarios, the dataset must contain actions from first-person perspective as well as third-person perspective. For this purpose, we introduce a new dataset which contains actions from both the perspectives captured using a head-mounted camera. We employ a motion pyramidal structure for grouping the dense trajectory features. The relative strengths of motion along the trajectories are used to compute different bag-of-words descriptors and concatenated to form a single descriptor for the action. The motion pyramidal approach performs better than the baseline improved trajectory descriptors. The method achieves 96.7% on the JPL interaction dataset and 61.8% on our NUS interaction dataset. The same is used to detect actions in long video sequences and achieves average precision of 0.79 on JPL interaction dataset.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "748260579dc2fb789335a88ae3f63c114795d047",
"rank": 1,
"score": 129255
},
{
"content": "Title: Action recognition with trajectory-pooled deep-convolutional descriptors Content: Visual features are of vital importance for human action understanding in videos. This paper presents a new video representation, called trajectory-pooled deep-convolutional descriptor (TDD), which shares the merits of both hand-crafted features [31] and deep-learned features [24]. Specifically, we utilize deep architectures to learn discriminative convolutional feature maps, and conduct trajectory-constrained pooling to aggregate these convolutional features into effective descriptors. To enhance the robustness of TDDs, we design two normalization methods to transform convolutional feature maps, namely spatiotemporal normalization and channel normalization. The advantages of our features come from (i) TDDs are automatically learned and contain high discriminative capacity compared with those hand-crafted features; (ii) TDDs take account of the intrinsic characteristics of temporal dimension and introduce the strategies of trajectory-constrained sampling and pooling for aggregating deep-learned features. We conduct experiments on two challenging datasets: HMD-B51 and UCF101. Experimental results show that TDDs outperform previous hand-crafted features [31] and deep-learned features [24]. Our method also achieves superior performance to the state of the art on these datasets.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "c6241e6fc94192df2380d178c4c96cf071e7a3ac",
"rank": 2,
"score": 123229
},
{
"content": "Title: Learning Compact Appearance Representation for Video-based Person Re-Identification Content: This paper presents a novel approach for video-based person re-identification using multiple Convolutional Neural Networks (CNNs). Unlike previous work, we intend to extract a compact yet discriminative appearance representation from several frames rather than the whole sequence. Specifically, given a video, the representative frames are selected based on the walking profile of consecutive frames. A multiple CNN architecture incorporated with feature pooling is proposed to learn and compile the features of the selected representative frames into a compact description about the pedestrian for identification. Experiments are conducted on benchmark datasets to demonstrate the superiority of the proposed method over existing person re-identification approaches.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "8552f6e3f73db564a2e625cceb1d1348d70b598c",
"rank": 3,
"score": 122402
},
{
"content": "Title: Rank Pooling for Action Recognition Content: We propose a function-based temporal pooling method that captures the latent structure of the video sequence data - e.g., how frame-level features evolve over time in a video. We show how the parameters of a function that has been fit to the video data can serve as a robust new video representation. As a specific example, we learn a pooling function via ranking machines. By learning to rank the frame-level features of a video in chronological order, we obtain a new representation that captures the video-wide temporal dynamics of a video, suitable for action recognition. Other than ranking functions, we explore different parametric models that could also explain the temporal changes in videos. The proposed functional pooling methods, and rank pooling in particular, is easy to interpret and implement, fast to compute and effective in recognizing a wide variety of actions. We evaluate our method on various benchmarks for generic action, fine-grained action and gesture recognition. Results show that rank pooling brings an absolute improvement of 7-10 average pooling baseline. At the same time, rank pooling is compatible with and complementary to several appearance and local motion based methods and features, such as improved trajectories and deep learning features.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "6674729287f2482eda9e836846d2a35e63ea401c",
"rank": 4,
"score": 114953
},
{
"content": "Title: First-Person Activity Recognition: What Are They Doing to Me? Content: This paper discusses the problem of recognizing interaction-level human activities from a first-person viewpoint. The goal is to enable an observer (e.g., a robot or a wearable camera) to understand 'what activity others are performing to it' from continuous video inputs. These include friendly interactions such as 'a person hugging the observer' as well as hostile interactions like 'punching the observer' or 'throwing objects to the observer', whose videos involve a large amount of camera ego-motion caused by physical interactions. The paper investigates multi-channel kernels to integrate global and local motion information, and presents a new activity learning/recognition methodology that explicitly considers temporal structures displayed in first-person activity videos. In our experiments, we not only show classification results with segmented videos, but also confirm that our new approach is able to detect activities from continuous videos reliably.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "1aad2da473888cb7ebc1bfaa15bfa0f1502ce005",
"rank": 5,
"score": 112673
},
{
"content": "Title: Boosted key-frame selection and correlated pyramidal motion-feature representation for human action recognition Content: In this paper we propose a novel method for human action recognition based on boosted key-frame selection and correlated pyramidal motion feature representations. Instead of using an unsupervised method to detect interest points, a Pyramidal Motion Feature (PMF), which combines optical flow with a biologically inspired feature, is extracted from each frame of a video sequence. The AdaBoost learning algorithm is then applied to select the most discriminative frames from a large feature pool. In this way, we obtain the top-ranked boosted frames of each video sequence as the key frames which carry the most representative motion information. Furthermore, we utilise the correlogram which focuses not only on probabilistic distributions within one frame but also on the temporal relationships of the action sequence. In the classification phase, a Support-Vector Machine (SVM) is adopted as the final classifier for human action recognition. To demonstrate generalizability, our method has been systematically tested on a variety of datasets and shown to be more effective and accurate for action recognition compared to the previous work. We obtain overall accuracies of: 95.5%, 93.7%, and 36.5% with our proposed method on the KTH, the multiview IXMAS and the challenging HMDB51 datasets, respectively. & 2012 Elsevier Ltd. All rights reserved.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "926b99cbea04fb213c9984b10acf2235a3949ebb",
"rank": 6,
"score": 107923
},
{
"content": "Title: ActionVLAD: Learning Spatio-Temporal Aggregation for Action Classification Content: In this work, we introduce a new video representation for action classification that aggregates local convolutional features across the entire spatio-temporal extent of the video. We do so by integrating state-of-the-art two-stream networks [42] with learnable spatio-temporal feature aggregation [6]. The resulting architecture is end-to-end trainable for whole-video classification. We investigate different strategies for pooling across space and time and combining signals from the different streams. We find that: (i) it is important to pool jointly across space and time, but (ii) appearance and motion streams are best aggregated into their own separate representations. Finally, we show that our representation outperforms the two-stream base architecture by a large margin (13% relative) as well as outperforms other baselines with comparable base architectures on HMDB51, UCF101, and Charades video classification benchmarks.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "63213d080a43660ac59ea12e3c35e6953f6d7ce8",
"rank": 7,
"score": 107464
},
{
"content": "Title: Higher-Order Pooling of CNN Features via Kernel Linearization for Action Recognition Content: Most successful deep learning algorithms for action recognition extend models designed for image-based tasks such as object recognition to video. Such extensions are typically trained for actions on single video frames or very short clips, and then their predictions from sliding-windows over the video sequence are pooled for recognizing the action at the sequence level. Usually this pooling step uses the first-order statistics of frame-level action predictions. In this paper, we explore the advantages of using higherorder correlations, specifically, we introduce Higher-order Kernel (HOK) descriptors generated from the late fusion of CNN classifier scores from all the frames in a sequence. To generate these descriptors, we use the idea of kernel linearization. Specifically, a similarity kernel matrix, which captures the temporal evolution of deep classifier scores, is first linearized into kernel feature maps. The HOK descriptors are then generated from the higher-order cooccurrences of these feature maps, and are then used as input to a video-level classifier. We provide experiments on two fine-grained action recognition datasets, and show that our scheme leads to state-of-the-art results.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "259817ee2a4419795f698b123027aad89cf5f903",
"rank": 8,
"score": 106544
},
{
"content": "Title: Point & Teleport Locomotion Technique for Virtual Reality Content: With the increasing popularity of virtual reality (VR) and new devices getting available with relatively lower costs, more and more video games have been developed recently. Most of these games use first person interaction techniques since it is more natural for Head Mounted Displays (HMDs). One of the most widely used interaction technique in VR video games is locomotion that is used to move user's viewpoint in virtual environments. Locomotion is an important component of video games since it can have a strong influence on user experience. In this study, a new locomotion technique we called \"Point & Teleport\" is described and compared with two commonly used VR locomotion techniques of walk-in-place and joystick. In this technique, users simply point where they want to be in virtual world and they are teleported to that position. As a major advantage, it is not expected to introduce motion sickness since it does not involve any visible translational motion. In this study, two VR experiments were designed and performed to analyze the Point & Teleport technique. In the first experiment, Point & Teleport was compared with walk-in-place and joystick locomotion techniques. In the second experiment, a direction component was added to the Point & Teleport technique so that the users could specify their desired orientation as well. 16 users took part in both experiments. Results indicated that Point & Teleport is a fun and user friendly locomotion method whereas the additional direction component degraded the user experience.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "e352612f51ad34c764f128cc62e91b51fe7a9759",
"rank": 9,
"score": 100871
},
{
"content": "Title: SIFT Features Tracking for Video Stabilization Content: This paper presents a video stabilization algorithm based on the extraction and tracking of scale invariant feature transform features through video frames. Implementation of SIFT operator is analyzed and adapted to be used in a feature-based motion estimation algorithm. SIFT features are extracted from video frames and then their trajectory is evaluated to estimate interframe motion. A modified version of iterative least squares method is adopted to avoid estimation errors and features are tracked as they appear in nearby frames to improve video stability. Intentional camera motion is eventually filtered with adaptive motion vector integration. Results confirm the effectiveness of the method.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "6ec17e735cd9f7cb37485ab07b905a7895b0067d",
"rank": 10,
"score": 98877
},
{
"content": "Title: Social interactions: A first-person perspective Content: This paper presents a method for the detection and recognition of social interactions in a day-long first-person video of u social event, like a trip to an amusement park. The location and orientation of faces are estimated and used to compute the line of sight for each face. The context provided by all the faces in a frame is used to convert the lines of sight into locations in space to which individuals attend. Further, individuals are assigned roles based on their patterns of attention. The rotes and locations of individuals are analyzed over time to detect and recognize the types of social interactions. In addition to patterns of face locations and attention, the head movements of the first-person can provide additional useful cues as to their attentional focus. We demonstrate encouraging results on detection and recognition of social interactions in first-person videos captured from multiple days of experience in amusement parks.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "014e1186209e4f942f3b5ba29b6b039c8e99ad88",
"rank": 11,
"score": 98874
},
{
"content": "Title: First-person hyper-lapse videos Content: We present a method for converting first-person videos, for example, captured with a helmet camera during activities such as rock climbing or bicycling, into hyper-lapse videos, i.e., time-lapse videos with a smoothly moving camera. At high speed-up rates, simple frame sub-sampling coupled with existing video stabilization methods does not work, because the erratic camera shake present in first-person videos is amplified by the speed-up. Our algorithm first reconstructs the 3D input camera path as well as dense, per-frame proxy geometries. We then optimize a novel camera path for the output video that passes near the input cameras while ensuring that the virtual camera looks in directions that can be rendered well from the input. Finally, we generate the novel smoothed, time-lapse video by rendering, stitching, and blending appropriately selected source frames for each output frame. We present a number of results for challenging videos that cannot be processed using traditional techniques.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "c67192cb7c82d2a0516b656909985823a5b2aba0",
"rank": 12,
"score": 98064
},
{
"content": "Title: Deep motion features for visual tracking Content: Robust visual tracking is a challenging computer vision problem, with many real-world applications. Most existing approaches employ hand-crafted appearance features, such as HOG or Color Names. Recently, deep RGB features extracted from convolutional neural networks have been successfully applied for tracking. Despite their success, these features only capture appearance information. On the other hand, motion cues provide discriminative and complementary information that can improve tracking performance. Contrary to visual tracking, deep motion features have been successfully applied for action recognition and video classification tasks. Typically, the motion features are learned by training a CNN on optical flow images extracted from large amounts of labeled videos. This paper presents an investigation of the impact of deep motion features in a tracking-by-detection framework. We further show that hand-crafted, deep RGB, and deep motion features contain complementary information. To the best of our knowledge, we are the first to propose fusing appearance information with deep motion features for visual tracking. Comprehensive experiments clearly suggest that our fusion approach with deep motion features outperforms standard methods relying on appearance information alone.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "264a84f4d27cd4bca94270620907cffcb889075c",
"rank": 13,
"score": 97661
},
{
"content": "Title: Self-organizing neural integration of pose-motion features for human action recognition Content: The visual recognition of complex, articulated human movements is fundamental for a wide range of artificial systems oriented toward human-robot communication, action classification, and action-driven perception. These challenging tasks may generally involve the processing of a huge amount of visual information and learning-based mechanisms for generalizing a set of training actions and classifying new samples. To operate in natural environments, a crucial property is the efficient and robust recognition of actions, also under noisy conditions caused by, for instance, systematic sensor errors and temporarily occluded persons. Studies of the mammalian visual system and its outperforming ability to process biological motion information suggest separate neural pathways for the distinct processing of pose and motion features at multiple levels and the subsequent integration of these visual cues for action perception. We present a neurobiologically-motivated approach to achieve noise-tolerant action recognition in real time. Our model consists of self-organizing Growing When Required (GWR) networks that obtain progressively generalized representations of sensory inputs and learn inherent spatio-temporal dependencies. During the training, the GWR networks dynamically change their topological structure to better match the input space. We first extract pose and motion features from video sequences and then cluster actions in terms of prototypical pose-motion trajectories. Multi-cue trajectories from matching action frames are subsequently combined to provide action dynamics in the joint feature space. Reported experiments show that our approach outperforms previous results on a dataset of full-body actions captured with a depth sensor, and ranks among the best results for a public benchmark of domestic daily actions.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "b4b3caf8e55aad64a31a29000ee990f93bf1755c",
"rank": 14,
"score": 96038
},
{
"content": "Title: Visual odometry for ground vehicle applications Content: We present a system that estimates the motion of a stereo head or a single moving camera based on video input. The system operates in real-time with low delay and the motion estimates are used for navigational purposes. The front end of the system is a feature tracker. Point features are matched between pairs of frames and linked into image trajectories at video rate. Robust estimates of the camera motion are then produced from the feature tracks using a geometric hypothesize-and-test architecture. This generates motion estimates from visual input alone. No prior knowledge of the scene nor the motion is necessary. The visual estimates can also be used in conjunction with information from other sources such as GPS, inertia sensors, wheel encoders, etc. The pose estimation method has been applied successfully to video from aerial, automotive and handheld platforms. We focus on results obtained with a stereo-head mounted on an autonomous ground vehicle. We give examples of camera trajectories estimated in real-time purely from images over previously unseen distances (600 meters) and periods of time .",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "c8965cc5c62a245593dbc679aebdf3338bb945fc",
"rank": 15,
"score": 95719
},
{
"content": "Title: Human Action Recognition Using Multi-Velocity STIPs and Motion Energy Orientation Histogram Content: Local image features in space-time or spatio-temporal interest points provide compact and abstract representations of patterns in a video sequence. In this paper, we present a novel human action recognition method based on multi-velocity spatio-temporal interest points (MVSTIPs) and a novel local descriptor called motion energy (ME) orientation histogram (MEOH). The MVSTIP detection includes three steps: first, filtering video frames with multi-direction ME filters at different speeds to detect significant changes at the pixel level; thereafter, a surround suppression model is employed to rectify the ME deviation caused by the camera motion and complicated backgrounds (e.g., dynamic texture); finally, MVSTIPs are obtained with local maximum filters at multispeeds. After detection, we develop MEOH descriptor to capture the motion features in local regions around interest points. The performance of the proposed method is evaluated on KTH, Weizmann, and UCF sports human action datasets. Results show that our method is robust to both simple and complex backgrounds and the method is superior to other methods that are based on local features.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "6985a4cb132b23778cd74bd9abed8764d103ae59",
"rank": 16,
"score": 94324
},
{
"content": "Title: Action Recognition with Improved Trajectories Content: Recently dense trajectories were shown to be an efficient video representation for action recognition and achieved state-of-the-art results on a variety of datasets. This paper improves their performance by taking into account camera motion to correct them. To estimate camera motion, we match feature points between frames using SURF descriptors and dense optical flow, which are shown to be complementary. These matches are, then, used to robustly estimate a homography with RANSAC. Human motion is in general different from camera motion and generates inconsistent matches. To improve the estimation, a human detector is employed to remove these matches. Given the estimated camera motion, we remove trajectories consistent with it. We also use this estimation to cancel out camera motion from the optical flow. This significantly improves motion-based descriptors, such as HOF and MBH. Experimental results on four challenging action datasets (i.e., Hollywood2, HMDB51, Olympic Sports and UCF50) significantly outperform the current state of the art.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "070874b011f8eb2b18c8aa521ad0a7a932b4d9ad",
"rank": 17,
"score": 93054
},
{
"content": "Title: Vehicle Detection and Tracking in Car Video Based on Motion Model Content: This paper aims at real-time in-car video analysis to detect and track vehicles ahead for safety, autodriving, and target tracing. This paper describes a comprehensive approach to localizing target vehicles in video under various environmental conditions. The extracted geometry features from the video are continuously projected onto a 1-D profile and are constantly tracked. We rely on temporal information of features and their motion behaviors for vehicle identification, which compensates for the complexity in recognizing vehicle shapes, colors, and types. We probabilistically model the motion in the field of view according to the scene characteristic and the vehicle motion model. The hidden Markov model (HMM) is used to separate target vehicles from the background and track them probabilistically. We have investigated videos of day and night on different types of roads, showing that our approach is robust and effective in dealing with changes in environment and illumination and that real-time processing becomes possible for vehicle-borne cameras.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "29374eed47527cdaf14aa55fdb3935fc2de78c96",
"rank": 18,
"score": 92774
},
{
"content": "Title: Task-Driven Feature Pooling for Image Classification Content: Feature pooling is an important strategy to achieve high performance in image classification. However, most pooling methods are unsupervised and heuristic. In this paper, we propose a novel task-driven pooling (TDP) model to directly learn the pooled representation from data in a discriminative manner. Different from the traditional methods (e.g., average and max pooling), TDP is an implicit pooling method which elegantly integrates the learning of representations into the given classification task. The optimization of TDP can equalize the similarities between the descriptors and the learned representation, and maximize the classification accuracy. TDP can be combined with the traditional BoW models (coding vectors) or the recent state-of-the-art CNN models (feature maps) to achieve a much better pooled representation. Furthermore, a self-training mechanism is used to generate the TDP representation for a new test image. A multi-task extension of TDP is also proposed to further improve the performance. Experiments on three databases (Flower-17, Indoor-67 and Caltech-101) well validate the effectiveness of our models.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "af390f5ded307d8f548243163178d2db639b6e5c",
"rank": 19,
"score": 92535
},
{
"content": "Title: Human detection in surveillance videos and its applications - a review Content: Detecting human beings accurately in a visual surveillance system is crucial for diverse application areas including abnormal event detection, human gait characterization, congestion analysis, person identification, gender classification and fall detection for elderly people. The first step of the detection process is to detect an object which is in motion. Object detection could be performed using background subtraction, optical flow and spatio-temporal filtering techniques. Once detected, a moving object could be classified as a human being using shape-based, texture-based or motion-based features. A comprehensive review with comparisons on available techniques for detecting human beings in surveillance videos is presented in this paper. The characteristics of few benchmark datasets as well as the future research directions on human detection have also been discussed.",
"qid": "0229829e9a1eed5769a2b5eccddcaa7cd9460b92",
"docid": "1f4fff64adef5ec6ae21e8647d5a042bf71d64d9",
"rank": 20,
"score": 92037
}
] |
General transformations for GPU execution of tree traversals
|
[
{
"content": "Title: KD-tree acceleration structures for a GPU raytracer Content: Modern graphics hardware architectures excel at compute-intensive tasks such as ray-triangle intersection, making them attractive target platforms for raytracing. To date, most GPU-based raytracers have relied upon uniform grid acceleration structures. In contrast, the kd-tree has gained widespread use in CPU-based raytracers and is regarded as the best general-purpose acceleration structure. We demonstrate two kd-tree traversal algorithms suitable for GPU implementation and integrate them into a streaming raytracer. We show that for scenes with many objects at different scales, our kd-tree algorithms are up to 8 times faster than a uniform grid. In addition, we identify load balancing and input data recirculation as two fundamental sources of inefficiency when raytracing on current graphics hardware.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "36f06481eaae63522dfb61475602584997ebfee8",
"rank": 1,
"score": 139538
},
{
"content": "Title: Path Reducing Watershed for the GPU Content: The watershed transform is a popular image segmentation procedure from mathematical morphology used in many applications of computer vision. This paper proposes a novel parallel watershed procedure designed for GPU implementation. Our algorithm constructs paths of steepest descent and reduces these paths into direct pointers to catchment basin minima in logarithmic time, also crucially incorporating successful resolution of plateaux. Three implementation variants and their parameters are analysed through experiments on 2D and 3D images; a comparison against the state-of-the-art shows a runtime improvement of around 30%. For 3D images of 128 megavoxels execution times of approximately 1.5–2 seconds are achieved.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "35bef4597f5e514359ff45bea31be8b8239effe1",
"rank": 2,
"score": 117693
},
{
"content": "Title: Parallel tree-ensemble algorithms for GPUs using CUDA Content: We present two new parallel implementations of the tree-ensemble algorithms Random Forest (RF) and Extremely randomized trees (ERT) for emerging many-core platforms, e.g., contemporary graphics cards suitable for general-purpose computing (GPGPU). Random Forest and Extremely randomized trees are ensemble learners for classification and regression. They operate by constructing a multitude of decision trees at training time and outputting a prediction by comparing the outputs of the individual trees. Thanks to the inherent parallelism of the task, an obvious platform for its computation is to employ contemporary GPUs with a large number of processing cores. Previous parallel algorithms for Random Forests in the literature are either designed for traditional multi-core CPU platforms or early history GPUs with simpler hardware architecture and relatively few number of cores. The new parallel algorithms are designed for contemporary GPUs with a large number of cores and take into account aspects of the newer hardware architectures as memory hierarchy and thread scheduling. They are implemented using the C/C++ language and the CUDA interface for best possible performance on NVidia-based GPUs. An experimental study comparing with the most important previous solutions for CPU and GPU platforms shows significant improvement for the new implementations, often with several magnitudes.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "698b8181cd613a72adeac0d75252afe7f57a5180",
"rank": 3,
"score": 114863
},
{
"content": "Title: FAST: fast architecture sensitive tree search on modern CPUs and GPUs Content: In-memory tree structured index search is a fundamental database operation. Modern processors provide tremendous computing power by integrating multiple cores, each with wide vector units. There has been much work to exploit modern processor architectures for database primitives like scan, sort, join and aggregation. However, unlike other primitives, tree search presents significant challenges due to irregular and unpredictable data accesses in tree traversal. In this paper, we present FAST, an extremely fast architecture sensitive layout of the index tree. FAST is a binary tree logically organized to optimize for architecture features like page size, cache line size, and SIMD width of the underlying hardware. FAST eliminates impact of memory latency, and exploits thread-level and datalevel parallelism on both CPUs and GPUs to achieve 50 million (CPU) and 85 million (GPU) queries per second, 5X (CPU) and 1.7X (GPU) faster than the best previously reported performance on the same architectures. FAST supports efficient bulk updates by rebuilding index trees in less than 0.1 seconds for datasets as large as 64Mkeys and naturally integrates compression techniques, overcoming the memory bandwidth bottleneck and achieving a 6X performance improvement over uncompressed index search for large keys on CPUs.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "295521cfe1a56458d53a58613de5fb92c97c5c23",
"rank": 4,
"score": 113494
},
{
"content": "Title: IMGPU: GPU-Accelerated Influence Maximization in Large-Scale Social Networks Content: Influence Maximization aims to find the top-$(K)$ influential individuals to maximize the influence spread within a social network, which remains an important yet challenging problem. Proven to be NP-hard, the influence maximization problem attracts tremendous studies. Though there exist basic greedy algorithms which may provide good approximation to optimal result, they mainly suffer from low computational efficiency and excessively long execution time, limiting the application to large-scale social networks. In this paper, we present IMGPU, a novel framework to accelerate the influence maximization by leveraging the parallel processing capability of graphics processing unit (GPU). We first improve the existing greedy algorithms and design a bottom-up traversal algorithm with GPU implementation, which contains inherent parallelism. To best fit the proposed influence maximization algorithm with the GPU architecture, we further develop an adaptive K-level combination method to maximize the parallelism and reorganize the influence graph to minimize the potential divergence. We carry out comprehensive experiments with both real-world and sythetic social network traces and demonstrate that with IMGPU framework, we are able to outperform the state-of-the-art influence maximization algorithm up to a factor of 60, and show potential to scale up to extraordinarily large-scale networks.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "38a08fbe5eabbd68db495fa38f4ee506d82095d4",
"rank": 5,
"score": 109292
},
{
"content": "Title: A GPU-Based Rasterization Algorithm for Boolean Operations on Polygons Content: This paper presents a new GPU-based rasterization algorithm for Boolean operations that handles arbitary closed polygons. We construct an efficient data structure for interoperation of CPU and GPU and propose a fast GPU-based contour extraction method to ensure the performance of our algorithm. We then design a novel traversing strategy to achieve an error-free calculation of intersection point for correct Boolean operations. We finally give a detail evaluation and the results show that our algorithm has a higher performance than exsiting algorithms on processing polygons with large amount of vertices. key words: GPU, CPU, rasterization, Boolean operation, error-free",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "67078d516a85204c016846e30c02e901ac16f142",
"rank": 6,
"score": 105951
},
{
"content": "Title: Kd-Jump: a Path-Preserving Stackless Traversal for Faster Isosurface Raytracing on GPUs Content: Stackless traversal techniques are often used to circumvent memory bottlenecks by avoiding a stack and replacing return traversal with extra computation. This paper addresses whether the stackless traversal approaches are useful on newer hardware and technology (such as CUDA). To this end, we present a novel stackless approach for implicit kd-trees, which exploits the benefits of index-based node traversal, without incurring extra node visitation. This approach, which we term Kd-Jump, enables the traversal to immediately return to the next valid node, like a stack, without incurring extra node visitation (kd-restart). Also, Kd-Jump does not require global memory (stack) at all and only requires a small matrix in fast constant-memory. We report that Kd-Jump outperforms a stack by 10 to 20% and kd-restar t by 100%. We also present a Hybrid Kd-Jump, which utilizes a volume stepper for leaf testing and a run-time depth threshold to define where kd-tree traversal stops and volume-stepping occurs. By using both methods, we gain the benefits of empty space removal, fast texture-caching and realtime ability to determine the best threshold for current isosurface and view direction.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "25eb08e6985ded20ae723ec668014a2bad789e0f",
"rank": 7,
"score": 104143
},
{
"content": "Title: A fast and efficient sift detector using the mobile GPU Content: Emerging mobile applications, such as augmented reality, demand robust feature detection at high frame rates. We present an implementation of the popular Scale-Invariant Feature Transform (SIFT) feature detection algorithm that incorporates the powerful graphics processing unit (GPU) in mobile devices. Where the usual GPU methods are inefficient on mobile hardware, we propose a heterogeneous dataflow scheme. By methodically partitioning the computation, compressing the data for memory transfers, and taking into account the unique challenges that arise out of the mobile GPU, we are able to achieve a speedup of 4-7x over an optimized CPU version, and a 6.4x speedup over a published GPU implementation. Additionally, we reduce energy consumption by 87 percent per image. We achieve near-realtime detection without compromising the original algorithm.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "114e15d677d2055a3a627eb805c2e914f405404e",
"rank": 8,
"score": 100855
},
{
"content": "Title: Small Discrete Fourier Transforms on GPUs Content: Efficient implementations of the Discrete Fourier Transform (DFT) for GPUs provide good performance with large data sizes, but are not competitive with CPU code for small data sizes. On the other hand, several applications perform multiple DFTs on small data sizes. In fact, even algorithms for large data sizes use a divide-and-conquer approach, where eventually small DFTs need to be performed. We discuss our DFT implementation, which is efficient for multiple small DFTs. One feature of our implementation is the use of the asymptotically slow matrix multiplication approach for small data sizes, which improves performance on the GPU due to its regular memory access and computational patterns. We combine this algorithm with the mixed radix algorithm for 1-D, 2-D, and 3-D complex DFTs. We also demonstrate the effect of different optimization techniques. When GPUs are used to accelerate a component of an application running on the host, it is important that decisions taken to optimize the GPU performance not affect the performance of the rest of the application on the host. One feature of our implementation is that we use a data layout that is not optimal for the GPU so that the overall effect on the application is better. Our implementation performs up to two orders of magnitude faster than cuFFT on an NVIDIA GeForce 9800 GTX GPU and up to one to two orders of magnitude faster than FFTW on a CPU for multiple small DFTs. Furthermore, we show that our implementation can accelerate the performance of a Quantum Monte Carlo application for which cuFFT is not effective. The primary contributions of this work lie in demonstrating the utility of the matrix multiplication approach and also in providing an implementation that is efficient for small DFTs when a GPU is used to accelerate an application running on the host.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "3e15fb211f94b2e85d07e60f0aa641797b062e01",
"rank": 9,
"score": 97139
},
{
"content": "Title: Attention Is All You Need Content: The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 Englishto-German translation task, improving over the existing best results, including ensembles, by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.2 after training for 4.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "0b0cf7e00e7532e38238a9164f0a8db2574be2ea",
"rank": 10,
"score": 96262
},
{
"content": "Title: Automatic C-to-CUDA Code Generation for Affine Programs Content: Graphics Processing Units (GPUs) offer tremendous computational power. CUDA (Compute Unified Device Architecture) provides a multi-threaded parallel programming model, facilitating high performance implementations of general-purpose computations. However, the explicitly managed memory hierarchy and multi-level parallel view make manual development of high-performance CUDA code rather complicated. Hence the automatic transformation of sequential input programs into efficient parallel CUDA programs is of considerable interest. This paper describes an automatic code transformation system that generates parallel CUDA code from input sequential C code, for regular (affine) programs. Using and adapting publicly available tools that have made polyhedral compiler optimization practically effective, we develop a C-to-CUDA transformation system that generates two-level parallel CUDA code that is optimized for efficient data access. The performance of automatically generated code is compared with manually optimized CUDA code for a number of benchmarks. The performance of the automatically generated CUDA code is quite close to hand-optimized CUDA code and considerably better than the benchmarks’ performance on a multicore CPU.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "bbe0f0b3e2d60c4f96d9d84f97dc8a9be4f72802",
"rank": 11,
"score": 95225
},
{
"content": "Title: Parallel Processing of Dynamic Continuous Queries over Streaming Data Flows Content: More and more real-time applications need to handle dynamic continuous queries over streaming data of high density. Conventional data and query indexing approaches generally do not apply for excessive costs in either maintenance or space. Aiming at these problems, this study first proposes a new indexing structure by fusing an adaptive cell and KDB-tree, namely CKDB-tree. A cell-tree indexing approach has been developed on the basis of the CKDB-tree that supports dynamic continuous queries. The approach significantly reduces the space costs and scales well with the increasing data size. Towards providing a scalable solution to filtering massive steaming data, this study has explored the feasibility to utilize the contemporary general-purpose computing on the graphics processing unit (GPGPU). The CKDB-tree-based approach has been extended to operate on both the CPU (host) and the GPU (device). The GPGPU-aided approach performs query indexing on the host while perform streaming data filtering on the device in a massively parallel manner. The two heterogeneous tasks execute in parallel and the latency of streaming data transfer between the host and the device is hidden. The experimental results indicate that (1) CKDB-tree can reduce the space cost comparing to the cell-based indexing structure by 60 percent on average, (2) the approach upon the CKDB-tree outperforms the traditional counterparts upon the KDB-tree by 66, 75 and 79 percent in average for uniform, skewed and hyper-skewed data in terms of update costs, and (3) the GPGPU-aided approach greatly improves the approach upon the CKDB-tree with the support of only a single Kepler GPU, and it provides real-time filtering of streaming data with 2.5M data tuples per second. The massively parallel computing technology exhibits great potentials in streaming data monitoring.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "aac0e2533870637e425a5ea8d4807676cfc4d0aa",
"rank": 12,
"score": 95045
},
{
"content": "Title: Merkle Tree Traversal in Log Space and Time Content: We present a technique for Merkle tree traversal which requires only logarithmic space and time. For a tree with N nodes, our algorithm computes sequential tree leaves and authentication path data in time Log2(N) and space less than 3Log2(N), where the units of computation are hash function evaluations or leaf value computations, and the units of space are the number of node values stored. Relative to this algorithm, we show our bounds to be necessary and sufficient. This result is an asymptotic improvement over all other previous results (for example, measuring cost = space ∗ time). We also prove that the complexity of our algorithm is optimal: There can exist no Merkle tree traversal algorithm which consumes both less than O(Log2(N)) space and less than O(Log2(N)) time. Our algorithm is especially of practical interest when space efficiency is required, and can also enhance other traversal algorithms which relax space constraints to gain speed.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "342ca79a0ac2cc726bf31ffb4ce399821d3e2979",
"rank": 13,
"score": 94941
},
{
"content": "Title: Fractal Merkle Tree Representation and Traversal Content: We introduce a technique for traversal of Merkle trees, and propose an efficient algorithm that generates a sequence of leaves along with their associated authentication paths. For one choice of parameters, and a total of N leaves, our technique requires a worst-case computational effort of 2 log N/loglog N hash function evaluations per output, and a total storage capacity of less than 1.5 log N/loglog N hash values. This is a simultaneous improvement both in space and time complexity over any previously published algorithm.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "1794435f6b541109ee9ea812d80d5b9add95aacd",
"rank": 14,
"score": 93737
},
{
"content": "Title: Full-resolution interactive CPU volume rendering with coherent BVH traversal Content: We present an efficient method for volume rendering by raycasting on the CPU. We employ coherent packet traversal of an implicit bounding volume hierarchy, heuristically pruned using preintegrated transfer functions, to exploit empty or homogeneous space. We also detail SIMD optimizations for volumetric integration, trilinear interpolation, and gradient lighting. The resulting system performs well on low-end and laptop hardware, and can outperform out-of-core GPU methods by orders of magnitude when rendering large volumes without level-of-detail (LOD) on a workstation. We show that, while slower than GPU methods for low-resolution volumes, an optimized CPU renderer does not require LOD to achieve interactive performance on large data sets.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "f4c5f7bdf3f7ce924cd42f26d2a9eb97ab8da4a3",
"rank": 15,
"score": 93175
},
{
"content": "Title: Real-Time CPU-Based Large-Scale Three-Dimensional Mesh Reconstruction Content: In robotics, especially in this era of autonomous driving, mapping is one key ability of a robot to be able to navigate through an environment, localize on it, and analyze its traversability. To allow for real-time execution on constrained hardware, the map usually estimated by feature-based or semidense SLAM algorithms is a sparse point cloud; a richer and more complete representation of the environment is desirable. Existing dense mapping algorithms require extensive use of graphics processing unit (GPU) computing and they hardly scale to large environments; incremental algorithms from sparse points still represent an effective solution when light computational effort is needed and big sequences have to be processed in real time. In this letter, we improved and extended the state-of-the-art incremental manifold mesh algorithm proposed by Litvinov and Lhuillier and extended by Romanoni and Matteucci. While these algorithms do not reconstruct the map in real time and they embed points from SLAM or structure from motion only when their position is fixed, in this letter, we propose the first incremental algorithm able to reconstruct a manifold mesh in real time through single core CPU processing, which is also able to modify the mesh according to three-dimensional points updates from the underlying SLAM algorithm. We tested our algorithm against two state-of-the-art incremental mesh mapping systems on the KITTI dataset, and we showed that, while accuracy is comparable, our approach is able to reach real-time performances thanks to an order of magnitude speed-up.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "c62e4ef5b459827a13cc1f7015dbf33ea7677e06",
"rank": 16,
"score": 92953
},
{
"content": "Title: gNUFFTW : Auto-Tuning for High-Performance GPU-Accelerated Non-Uniform Fast Fourier Transforms by Teresa Content: Non-uniform sampling of the Fourier transform appears in many important applications such as magnetic resonance imaging (MRI), optics, tomography and radio interferometry. Computing the inverse often requires fast application of the non-uniform discrete Fourier transform (NUDFT) and its adjoint operation. Non-Uniform Fast Fourier Transform (NUFFT) methods, such as gridding/regridding, are approximate algorithms which often leverage the highly-optimized Fast Fourier Transform (FFT) and localized interpolations. These approaches require selecting several parameters, such as interpolation and FFT grid sizes, which affect both the accuracy and runtime. In addition, different implementations lie on a spectrum of precomputation levels, which can further speed up repeated computations, with various trade-offs in planning time, execution time and memory usage. Choosing the optimal parameters and implementations is important for performance speed, but difficult to do manually since the performance of NUFFT is not well-understood for modern parallel processors. Inspired by the FFTW library, we demonstrate an empirical auto-tuning approach for the NUFFT on General Purpose Graphics Processors Units (GPGPU). We demonstrate order-of-magnitude speed improvements with autotuning compared to typical default choices. Our auto-tuning is implemented in an easy to use proof-of-concept library called gNUFFTW, which leverages existing open-source NUFFT packages, cuFFT and cuSPARSE libraries, as well as our own NUFFT implementations for high performance. Keywords—non-uniform, non-Cartesian, FFT, NUFFT, GPU, auto-tuning, image processing.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "d1cf19c3f9d202d81cd67a999d7c831aa871cc6e",
"rank": 17,
"score": 92383
},
{
"content": "Title: Clause Restructuring for Statistical Machine Translation Content: We describe a method for incorporating syntactic information in statistical machine translation systems. The first step of the method is to parse the source language string that is being translated. The second step is to apply a series of transformations to the parse tree, effectively reordering the surface string on the source language side of the translation system. The goal of this step is to recover an underlying word order that is closer to the target language word-order than the original string. The reordering approach is applied as a pre-processing step in both the training and decoding phases of a phrase-based statistical MT system. We describe experiments on translation from German to English, showing an improvement from 25.2% Bleu score for a baseline system to 26.8% Bleu score for the system with reordering, a statistically significant improvement.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "555e1d6ecc7af031f29b0225bdca06d4a6da77ed",
"rank": 18,
"score": 90783
},
{
"content": "Title: Realtime Ray Tracing on GPU with BVH-based Packet Traversal Content: Recent GPU ray tracers can already achieve performance competitive to that of their CPU counterparts. Nevertheless, these systems can not yet fully exploit the capabilities of modern GPUs and can only handle medium-sized, static scenes. In this paper we present a BVH-based GPU ray tracer with a parallel packet traversal algorithm using a shared stack. We also present a fast, CPU-based BVH construction algorithm which very accurately approximates the surface area heuristic using streamed binning while still being one order of magnitude faster than previously published results. Furthermore, using a BVH allows us to push the size limit of supported scenes on the GPU: We can now ray trace the 12.7 million triangle Power Plant at 1024 times 1024 image resolution with 3 fps, including shading and shadows.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "826a530b835a917200da1b25993b5319021e4551",
"rank": 19,
"score": 90733
},
{
"content": "Title: Efficient Architecture Search by Network Transformation Content: Techniques for automatically designing deep neural network architectures such as reinforcement learning based approaches have recently shown promising results. However, their success is based on vast computational resources (e.g. hundreds of GPUs), making them difficult to be widely used. A noticeable limitation is that they still design and train each network from scratch during the exploration of the architecture space, which is highly inefficient. In this paper, we propose a new framework toward efficient architecture search by exploring the architecture space based on the current network and reusing its weights. We employ a reinforcement learning agent as the meta-controller, whose action is to grow the network depth or layer width with function-preserving transformations. As such, the previously validated networks can be reused for further exploration, thus saves a large amount of computational cost. We apply our method to explore the architecture space of the plain convolutional neural networks (no skip-connections, branching etc.) on image benchmark datasets (CIFAR-10, SVHN) with restricted computational resources (5 GPUs). Our method can design highly competitive networks that outperform existing networks using the same design scheme. On CIFAR-10, our model without skip-connections achieves 4.23% test error rate, exceeding a vast majority of modern architectures and approaching DenseNet. Furthermore, by applying our method to explore the DenseNet architecture space, we are able to achieve more accurate networks with fewer parameters.",
"qid": "027e7780dbda48d99f3654e77b4a63063224950e",
"docid": "84e65a5bdb735d62eef4f72c2f01af354b2285ba",
"rank": 20,
"score": 88435
}
] |
A Gentle Introduction to Soar, an Architecture for Human Cognition.
| [{"content":"Title: Ambiguity Resolution in a Cognitive Model of Language Comprehension Content: The(...TRUNCATED)
|
DCAN: Dual Channel-Wise Alignment Networks for Unsupervised Scene Adaptation
| [{"content":"Title: Unsupervised Visual Domain Adaptation Using Subspace Alignment Content: In this (...TRUNCATED)
|
AntNet: Distributed Stigmergetic Control for Communications Networks
| [{"content":"Title: A Survey of Recent Results in Networked Control Systems Content: Networked contr(...TRUNCATED)
|
Feature Extraction and Duplicate Detection for Text Mining : A Survey
| [{"content":"Title: Multimodal feature extraction and fusion for semantic mining of soccer video: a (...TRUNCATED)
|
End of preview. Expand
in Data Studio
No dataset card yet
- Downloads last month
- 6