HandBooster: Boosting 3D Hand-Mesh Reconstruction by Conditional Synthesis and Sampling of Hand-Object Interactions

TL;DR

HandBooster enhances 3D hand-mesh reconstruction from single images by using a conditional diffusion model to generate diverse, realistic hand-object interaction images, improving baseline performance on standard benchmarks.

Contribution

It introduces a novel conditional generative framework with a sampling strategy to synthesize diverse training data for better 3D hand reconstruction.

Findings

Significant performance improvements over SOTA on HO3D and DexYCB benchmarks.

Effective generation of diverse, realistic hand-object interaction images.

Enhanced 3D hand-mesh reconstruction accuracy from single images.

Abstract

Reconstructing 3D hand mesh robustly from a single image is very challenging, due to the lack of diversity in existing real-world datasets. While data synthesis helps relieve the issue, the syn-to-real gap still hinders its usage. In this work, we present HandBooster, a new approach to uplift the data diversity and boost the 3D hand-mesh reconstruction performance by training a conditional generative space on hand-object interactions and purposely sampling the space to synthesize effective data samples. First, we construct versatile content-aware conditions to guide a diffusion model to produce realistic images with diverse hand appearances, poses, views, and backgrounds; favorably, accurate 3D annotations are obtained for free. Then, we design a novel condition creator based on our similarity-aware distribution sampling strategies to deliberately find novel and realistic interaction poses that are distinctive from the training set. Equipped with our method, several baselines can be significantly improved beyond the SOTA on the HO3D and DexYCB benchmarks. Our code will be released on https://github.com/hxwork/HandBooster_Pytorch.