Reconstructing Hand-Object Interactions in the Wild

TL;DR

This paper introduces an optimization-based method for reconstructing 3D hand-object interactions in natural settings without requiring direct 3D supervision, leveraging various 2D and 3D data constraints.

Contribution

It presents a novel joint optimization approach that incorporates contact, collision, and occlusion constraints to improve 3D reconstruction in the wild without needing labeled 3D data.

Findings

Produces high-quality reconstructions on in-the-wild datasets

Outperforms existing methods in lab settings with ground truth 3D annotations

Effectively handles diverse object categories in natural environments

Abstract

In this work we explore reconstructing hand-object interactions in the wild. The core challenge of this problem is the lack of appropriate 3D labeled data. To overcome this issue, we propose an optimization-based procedure which does not require direct 3D supervision. The general strategy we adopt is to exploit all available related data (2D bounding boxes, 2D hand keypoints, 2D instance masks, 3D object models, 3D in-the-lab MoCap) to provide constraints for the 3D reconstruction. Rather than optimizing the hand and object individually, we optimize them jointly which allows us to impose additional constraints based on hand-object contact, collision, and occlusion. Our method produces compelling reconstructions on the challenging in-the-wild data from the EPIC Kitchens and the 100 Days of Hands datasets, across a range of object categories. Quantitatively, we demonstrate that our approach compares favorably to existing approaches in the lab settings where ground truth 3D annotations are available.