Hoodie Occlusion Image Dataset

The retail e-commerce industry is currently facing challenges with accurate product recognition due to occlusions in images, which can hinder customer experience and lead to lost sales. Existing solutions often struggle with occlusions as traditional image datasets do not adequately cover varied real-world scenarios. This dataset aims to address the issue of identifying hooded garments that are partially obscured by other objects, which is critical for improving automated recognition systems in online retail. The dataset is constructed from images collected in various shopping environments using high-resolution cameras, ensuring diverse lighting and angles. Quality control measures include multi-round annotations by trained professionals and consistency checks to ensure high accuracy. The images are stored in JPG format organized by categories of occlusion.