GaussianBody: Clothed Human Reconstruction via 3d Gaussian Splatting
This work solves the problem of reconstructing dynamic clothed humans from images for applications in computer graphics and VR, representing an incremental improvement over existing methods.
The paper tackles clothed human reconstruction by proposing GaussianBody, a method based on 3D Gaussian Splatting that addresses non-rigid deformations and cloth details, achieving state-of-the-art photorealistic novel-view rendering with explicit geometry reconstruction.
In this work, we propose a novel clothed human reconstruction method called GaussianBody, based on 3D Gaussian Splatting. Compared with the costly neural radiance based models, 3D Gaussian Splatting has recently demonstrated great performance in terms of training time and rendering quality. However, applying the static 3D Gaussian Splatting model to the dynamic human reconstruction problem is non-trivial due to complicated non-rigid deformations and rich cloth details. To address these challenges, our method considers explicit pose-guided deformation to associate dynamic Gaussians across the canonical space and the observation space, introducing a physically-based prior with regularized transformations helps mitigate ambiguity between the two spaces. During the training process, we further propose a pose refinement strategy to update the pose regression for compensating the inaccurate initial estimation and a split-with-scale mechanism to enhance the density of regressed point clouds. The experiments validate that our method can achieve state-of-the-art photorealistic novel-view rendering results with high-quality details for dynamic clothed human bodies, along with explicit geometry reconstruction.