Research map / 3D vision / Gaussian splatting
ContraGS: Codebook-Condensed and Trainable Gaussian Splatting for Fast, Memory-Efficient Reconstruction
Sankeerth Durvasula, Sharanshangar Muhunthan, Zain Moustafa, Richard Chen, Ruofan Liang, Yushi Guan, Nilesh Ahuja, Nilesh Jain, Selvakumar Panneer, Nandita Vijaykumar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| train on codebook-compressed Gaussians instead of full parameters (2M Gaussians) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | Peak model memory (MB): 473.000 -> 130.038 | Improved |
| train on codebook-compressed Gaussians (2M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | PSNR: 30.71 -> 30.06 | Hurt |
| train on codebook-compressed Gaussians (2M) | 3DGS-MCMC · Deep Blending and Tanks and Temples | PSNR: 33.99 -> 33.73; 24.32 -> 24.35 | No clear effect |
| train on codebook-compressed Gaussians (5M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | PSNR: 32.72 -> 31.01 | Hurt |
| train on codebook-compressed Gaussians (5M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | Peak model memory (MB): 1182.000 -> 275.594 | Improved |
| train on codebook-compressed Gaussians (2M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | Training speed (it/s): 23.65 -> 32.02 | Improved |
| train on codebook-compressed Gaussians (2M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | FPS: 102 -> 216 | Improved |
| train on codebook-compressed Gaussians (5M) | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | FPS: 65 -> 87 | Improved |
| raise Gaussian cap from 2M to 5M (ablation) | ContraGS · MipNeRF360 (7 scenes, avg) | PSNR: 30.06 -> 31.01 | Improved |
| use ContraGS-5M instead of MCMC-2M | 3DGS-MCMC · MipNeRF360 (7 scenes, avg) | PSNR: 30.71 -> 31.01 | Improved |
| use ContraGS-2M instead of 3DGS | 3DGS · MipNeRF360 (7 scenes, avg) | Peak model memory (MB): 2089.363 -> 130.038 | Improved |
| use ContraGS-2M instead of 3DGS (random init Blender) | 3DGS · Blender (8 scenes, random init) | PSNR: 31.07 -> 36.840 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Scaffold-GS — Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering (CVPR 2024) · better on 2 of 2 shared results
- EAGLES — EAGLES: Efficient Accelerated 3D Gaussians with Lightweight EncodingS (ECCV 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Optimized Minimal 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.92)
- FlexGS: Train Once, Deploy Everywhere with Many-in-One Flexible 3D Gaussian Splatting (CVPR 2025 · similarity 0.92)
- On Scaling Up 3D Gaussian Splatting Training (ICLR 2025 · similarity 0.92)
- DOGS: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus (NeurIPS 2024 · similarity 0.91)
- Fast Feedforward 3D Gaussian Splatting Compression (ICLR 2025 · similarity 0.91)
- ContextGS : Compact 3D Gaussian Splatting with Anchor Level Context Model (NeurIPS 2024 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
