MotionGS: Exploring Explicit Motion Guidance for Deformable 3D Gaussian Splatting
Preprint
Abstract
Dynamic scene reconstruction is a long-term challenge in the field of 3D vision. Recently, the emergence of 3D Gaussian Splatting has provided new insights into this problem. Although subsequent efforts rapidly extend static 3D Gaussian to dynamic scenes, they often lack explicit constraints on object motion, leading to optimization difficulties and performance degradation. To address the above issues, we propose a novel deformable 3D Gaussian splatting framework called MotionGS, which explores explicit motion priors to guide the deformation of 3D Gaussians. Specifically, we first introduce an optical flow decoupling module that decouples optical flow into camera flow and motion flow, corresponding to camera movement and object motion respectively. Then the motion flow can effectively constrain the deformation of 3D Gaussians, thus simulating the motion of dynamic objects. Additionally, a camera pose refinement module is proposed to alternately optimize 3D Gaussians and camera poses, mitigating the impact of inaccurate camera poses. Extensive experiments in the monocular dynamic scenes validate that MotionGS surpasses state-of-the-art methods and exhibits significant superiority in both qualitative and quantitative results. Project page: https://ruijiezhu94.github.io/MotionGS_pageKeywords
All Related Versions
This publication has 20 references indexed in Scilit:
- MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-view Images2024
- LGM: Large Multi-view Gaussian Model for High-Resolution 3D Content Creation2024
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting2024
- 3DGStream: On-the-Fly Training of 3D Gaussians for Efficient Streaming of Photo-Realistic Free-Viewpoint Videos2024
- Spacetime Gaussian Feature Splatting for Real-Time Dynamic View Synthesis2024
- SC-GS: Sparse-Controlled Gaussian Splatting for Editable Dynamic Scenes2024
- Neural Parametric Gaussians for Monocular Non-Rigid Object Reconstruction2024
- Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstruction2024
- 4K4D: Real-Time 4D View Synthesis at 4K Resolution2024
- Dynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis2024