Thick-soled Shoe Occlusion Image Dataset

The retail e-commerce sector is currently facing challenges in accurately recognizing products in images due to various occlusions and distortions. Existing solutions often struggle with partially obscured items or unfavorable shooting angles, leading to inconsistent recognition rates. This dataset aims to address these technical challenges by providing a rich set of images featuring thick-soled shoes under varying conditions of occlusion and interference. Data collection involved capturing images from multiple angles and under diverse lighting conditions using high-resolution cameras, ensuring a comprehensive dataset. Quality control measures included multi-round annotations, consistency checks by domain experts, and thorough reviews of each image to ensure accuracy. The dataset is organized in JPG format, with each image labeled according to its unique ID, occlusion level, angle, and any foreground objects present.