Women's Overcoat Occlusion Image Dataset

The retail e-commerce sector faces challenges in accurately identifying apparel items due to occlusion caused by background objects and model poses. Current solutions often rely on limited datasets that do not adequately represent diverse occlusion scenarios, leading to poor recognition performance. This dataset aims to address the need for a comprehensive collection of images exhibiting various occlusions to improve object detection and segmentation algorithms. The dataset comprises images collected from online fashion retailers, capturing different poses and occlusions. Quality control measures include multi-round annotations and expert reviews to ensure high accuracy and consistency. Data is stored in JPG format, organized by categories and tagged for easy access.