ArtVIP Machine Interactive Image Dataset

ArtVIP is a machine interactive image dataset released by the Beijing Humanoid Robot Innovation Center in 2025. The related paper results are "ArtVIP: Articulated Digital Assets of Visual Realism, Modular Interaction, and Physical Fidelity for Robot Learning".

The dataset contains 206 articulated objects in 26 categories, covering household items, large furniture, large appliances, small appliances and small furniture, and also provides 6 digital twin environments and 6 fully interactive environments, including children's rooms, dining rooms, kitchens, kitchens with living rooms, large living rooms and small living rooms. It ensures visual realism through precise geometric meshes and high-resolution textures, achieves physical fidelity through fine-tuned dynamic parameters, and is the first to embed modular interactive behaviors in assets and achieve pixel-level affordance annotation.