ShapeGaussian: High-Fidelity 4D Human Reconstruction in Monocular Videos via Vision Priors

TL;DR

ShapeGaussian is a novel method for high-fidelity 4D human reconstruction from monocular videos that combines vision priors with a two-step process to improve accuracy and robustness over existing template-based and template-free approaches.

Contribution

It introduces a template-free approach that integrates vision priors and a neural deformation model for superior 4D human reconstruction from monocular videos.

Findings

Outperforms template-based methods in accuracy and visual quality.

Effectively mitigates pose estimation errors using vision priors.

Handles diverse human motions robustly in casual videos.

Abstract

We introduce ShapeGaussian, a high-fidelity, template-free method for 4D human reconstruction from casual monocular videos. Generic reconstruction methods lacking robust vision priors, such as 4DGS, struggle to capture high-deformation human motion without multi-view cues. While template-based approaches, primarily relying on SMPL, such as HUGS, can produce photorealistic results, they are highly susceptible to errors in human pose estimation, often leading to unrealistic artifacts. In contrast, ShapeGaussian effectively integrates template-free vision priors to achieve both high-fidelity and robust scene reconstructions. Our method follows a two-step pipeline: first, we learn a coarse, deformable geometry using pretrained models that estimate data-driven priors, providing a foundation for reconstruction. Then, we refine this geometry using a neural deformation model to capture fine-grained dynamic details. By leveraging 2D vision priors, we mitigate artifacts from erroneous pose estimation in template-based methods and employ multiple reference frames to resolve the invisibility issue of 2D keypoints in a template-free manner. Extensive experiments demonstrate that ShapeGaussian surpasses template-based methods in reconstruction accuracy, achieving superior visual quality and robustness across diverse human motions in casual monocular videos.