4 datasets found

MANUS-Grasps is a large real-world multi-view RGB grasp dataset with over 7M frames from 53 cameras, providing full 360-degree coverage of 400+ grasps in over 30 diverse...

Two benchmark sets for measuring the extent to which language-and-vision language models use the visual signal in the presence or absence of stereotypes.

We aggregate 7 diverse real-world interaction datasets resulting in long-tailed collection of interactions across 157 object categories, and train a shared model across these.

We propose G-HOP, a denoising diffusion based generative prior for hand-object interactions that allows modeling both the 3D object and a human hand, conditioned on the object...