Predicting 4D Hand Trajectory from Monocular Videos
This work addresses the challenge of consistent 4D hand trajectory estimation for applications in human-computer interaction and robotics, representing an incremental advance by adapting existing methods to a new task.
The paper tackles the problem of predicting coherent 4D hand trajectories from monocular videos, which existing methods underperform on due to data scarcity, by repurposing an image-based transformer with novel attention layers, resulting in significant improvements in global trajectory accuracy while maintaining strong 2D reprojection alignment.
We present HaPTIC, an approach that infers coherent 4D hand trajectories from monocular videos. Current video-based hand pose reconstruction methods primarily focus on improving frame-wise 3D pose using adjacent frames rather than studying consistent 4D hand trajectories in space. Despite the additional temporal cues, they generally underperform compared to image-based methods due to the scarcity of annotated video data. To address these issues, we repurpose a state-of-the-art image-based transformer to take in multiple frames and directly predict a coherent trajectory. We introduce two types of lightweight attention layers: cross-view self-attention to fuse temporal information, and global cross-attention to bring in larger spatial context. Our method infers 4D hand trajectories similar to the ground truth while maintaining strong 2D reprojection alignment. We apply the method to both egocentric and allocentric videos. It significantly outperforms existing methods in global trajectory accuracy while being comparable to the state-of-the-art in single-image pose estimation. Project website: https://judyye.github.io/haptic-www