JonathonLuiten/Dynamic3DGaussians

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

2292 stars · 149 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official PyTorch implementation of 'Dynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis' (3DV 2024), which models dynamic 3D scenes as sets of 3D Gaussians that move and rotate over time. It extends Gaussian Splatting to dynamic scenes, enabling novel-view synthesis and dense 3D 6-DOF tracking, with training scripts and a custom visualizer.

Use cases

• reconstruct dynamic 3D scenes from multi-camera video
• track objects in 3D with 6-DOF trajectories
• render novel views of dynamic scenes
• extend gaussian splatting to time-varying scenes
• visualize pretrained dynamic 3D gaussian models
• research codebase for dynamic scene representation

When to choose

• you need dense 3D tracking plus novel-view synthesis for dynamic multi-view captures
• you want a clean, functional research codebase to build dynamic gaussian splatting on
• you have multi-camera video data and a CUDA GPU for training

When to avoid

• you only need static scene reconstruction (original Gaussian Splatting suffices)
• you lack multi-view video data or GPU resources
• you need a production-ready, licensed product (license is non-standard)

Facets

library · maturity maintenance

computer-vision graphics machine-learning simulation computer-vision graphics deep-learning python gaussian-splatting 3d-reconstruction novel-view-synthesis 6dof-tracking dynamic-scenes research-code pytorch 3dv-2024 linux gpu cuda

Member repositories

For agents

markdown · JSON · MCP: product_card(name="JonathonLuiten/Dynamic3DGaussians")