DeepFashion: Powering Robust Clothes Recognition and Retrieval with Rich Annotations

This paper introduces DeepFashion, a large-scale clothes database with comprehensive annotations. It addresses the challenges of recognizing and retrieving clothing items by providing a rich dataset for training robust models. The database includes a wide variety of clothing categories and styles, along with detailed attributes, allowing researchers to develop and evaluate algorithms for tasks such as fashion recognition, retrieval, and landmark localization.