Sweater Fitting Display Dataset

The current retail e-commerce industry faces a lack of virtual fitting experiences, especially in the apparel sector, where consumers often cannot fully understand the actual effect of products before purchasing. Existing virtual fitting solutions mostly rely on simple static image displays, lacking realism and interactivity, making it difficult for consumers to make satisfactory purchasing decisions. The sweater fitting display dataset aims to provide high-quality composite images to aid in developing more realistic virtual fitting systems. This dataset includes composite images of different models and sweater styles, catering to diverse user needs. Data collection employs advanced computer graphics technology, generating composite images through 3D modeling and rendering to ensure image quality. For quality control, we implement multiple rounds of labeling and consistency checks, with all images reviewed by experts to ensure annotation accuracy and consistency in the dataset. The final data is stored in JPEG format, organized by category and feature, facilitating subsequent use and expansion.