MovingFashion is the first publicly available benchmark designed to address the video-to-shop challenge in computer vision, where the goal is to retrieve fashion items worn in social media videos (e.g., Instagram, TikTok) by matching them to corresponding e-commerce product images.
GitHub Repository
license: cc-by-nc-4.0
Total Videos: 14,855 social videos
Source Platforms: Instagram, TikTok, and Net-A-Porter
Associated Shop… See the full description on the dataset page:
https://huggingface.co/datasets/christianjoppi/MovingFashion.
