Yoga Pants Occlusion Image Dataset

The current retail e-commerce industry faces challenges in accurately detecting product visibility due to occlusions caused by apparel styles, particularly yoga pants. Existing solutions often struggle with the variability of occlusion types and angles, leading to reduced accuracy in automated systems. This dataset aims to tackle the issue of identifying and classifying images where the waist or knee areas of yoga pants are obscured by motion blur or camera angles. The data is collected using high-resolution cameras in controlled environments with varying lighting to simulate real-world conditions. Quality control is performed through multiple annotation rounds, consistency checks, and expert reviews. The dataset is organized in JPG format, with each image accompanied by detailed annotations for effective training and evaluation.