BG-HOP: A Bimanual Generative Hand-Object Prior

TL;DR

BG-HOP is a generative model that captures 3D bimanual hand-object interactions, enabling synthesis of realistic grasps and interactions despite limited data, advancing 3D interaction modeling.

Contribution

It introduces BG-HOP, a novel generative prior specifically designed for modeling bimanual hand-object interactions in 3D, extending single-hand priors.

Findings

Successfully generates bimanual interactions.

Capable of synthesizing grasps for given objects.

Addresses data scarcity in bimanual interaction modeling.

Abstract

In this work, we present BG-HOP, a generative prior that seeks to model bimanual hand-object interactions in 3D. We address the challenge of limited bimanual interaction data by extending existing single-hand generative priors, demonstrating preliminary results in capturing the joint distribution of hands and objects. Our experiments showcase the model's capability to generate bimanual interactions and synthesize grasps for given objects. We make code and models publicly available.