WHOLE: World-Grounded Hand-Object Lifted from Egocentric Videos
This work addresses the problem of inconsistent and inaccurate hand-object pose estimation in egocentric videos for applications in robotics and AR/VR, representing a novel approach rather than an incremental improvement.
The paper tackles the problem of reconstructing hand and object motion from egocentric videos, which is challenging due to occlusions and objects moving out of view, by introducing WHOLE, a method that jointly reasons about hand-object interactions using a generative prior, achieving state-of-the-art performance in hand motion estimation, 6D object pose estimation, and interaction reconstruction.
Egocentric manipulation videos are highly challenging due to severe occlusions during interactions and frequent object entries and exits from the camera view as the person moves. Current methods typically focus on recovering either hand or object pose in isolation, but both struggle during interactions and fail to handle out-of-sight cases. Moreover, their independent predictions often lead to inconsistent hand-object relations. We introduce WHOLE, a method that holistically reconstructs hand and object motion in world space from egocentric videos given object templates. Our key insight is to learn a generative prior over hand-object motion to jointly reason about their interactions. At test time, the pretrained prior is guided to generate trajectories that conform to the video observations. This joint generative reconstruction substantially outperforms approaches that process hands and objects separately followed by post-processing. WHOLE achieves state-of-the-art performance on hand motion estimation, 6D object pose estimation, and their relative interaction reconstruction. Project website: https://judyye.github.io/whole-www