| # Dataset Description: This dataset contains comprehensive coffee bean package images specifically designed for Optical Character Recognition (OCR) tasks. | |
| The collection includes three distinct datasets: LRCB (Low-Resolution Coffee Bean), HRCB (High-Resolution Coffee Bean), and POIE (Product OCR Image Evaluation), providing researchers and practitioners with diverse image qualities and real-world scenarios for developing and evaluating OCR systems. | |
| # Dataset Composition: | |
| LRCB Dataset (Low-Resolution Coffee Bean) | |
| Images: 656 images | |
| Resolution Range: 480×640 to 800×600 pixels | |
| Source: Smartphone camera captures under various environmental conditions | |
| Purpose: Simulates typical consumer applications, mobile scanning scenarios, and low-end imaging devices | |
| # HRCB Dataset (High-Resolution Coffee Bean) | |
| Images: 503 high-quality images | |
| Resolution Range: 1920×1080 to 4000×3000 pixels | |
| Source: Professional cameras and high-end mobile devices | |
| Purpose: Represents scenarios with high-quality imaging equipment, professional product photography, and premium mobile captures | |
| # POIE Dataset (Product OCR Image Evaluation) [1] | |
| Type: Public validation dataset | |
| Purpose: Cross-domain evaluation and generalization testing | |
| Coverage: Diverse product categories beyond coffee packaging | |
| # References | |
| [1] Kuang, J.; Hua, W.; Liang, D.; Yang, M.; Jiang, D.; Ren, B.; Bai, X. Visual information extraction in the wild: Practical dataset and end-to-end solution. In {\em International Conference on Document Analysis and Recognition}; Springer: 2023; pp 36--53. | |