Research map / 3D vision / Gaussian splatting
Deformable Triangle Splatting: Flexible Primitives for Real-Time Radiance Field Rendering
Oriol Jiménez-Ayguadé, Antonio Agudo
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 |
|---|---|---|---|
| add learnable control-point deformation (ablation) | DETRIS · Mip-NeRF 360 + T&T (average) | PSNR: 26.500 -> 26.588 | Improved |
| add learnable control-point deformation (ablation) | DETRIS · Mip-NeRF 360 + T&T (average) | LPIPS: 0.180 -> 0.174 | Improved |
| add curvature regularizer (ablation) | DETRIS · Mip-NeRF 360 + T&T (average) | PSNR: 26.549 -> 26.588 | No clear effect |
| add curvature regularizer (ablation) | DETRIS · control point displacement | fraction |d|>0.1: 36.0% -> 29.0% | Improved |
| add control point warm-up delay (3000 iters) (ablation) | DETRIS · Mip-NeRF 360 + T&T (average) | PSNR: 26.545 -> 26.588 | Improved |
| increase K from 0 to 3 (ablation) | DETRIS · Bonsai | PSNR: 31.58 -> 32.44 | Improved |
| increase K from 0 to 3 (ablation) | DETRIS · Garden | PSNR: 27.08 -> 27.41 | Improved |
| increase K from 3 to 5 (ablation) | DETRIS · Bonsai | PSNR: 32.44 -> 32.52 | No clear effect |
| increase K from 0 to 3 (ablation) | DETRIS · Garden | FPS: 49 -> 33 | Hurt |
| increase K from 2 to 4 on crescent shape fitting (ablation) | DETRIS primitive · Crescent fit | L1: 0.0185 -> 0.0112 | Improved |
| use barycentric-space deformation instead of image-space window (ablation) | Triangle Splatting window · undeformed triangle rotation | max pixel discrepancy: 76 at 45 degrees, 181 at 60 degrees | Improved |
| replace rigid triangle with deformable triangle | Triangle Splatting · Mip-NeRF 360 (average) | LPIPS: 0.191 -> 0.183 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Zip-NeRF — Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fields (ICCV 2023) · 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
- 3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes (CVPR 2025 · similarity 0.93)
- Deformable Radial Kernel Splatting (CVPR 2025 · similarity 0.90)
- Sparse Voxels Rasterization: Real-time High-fidelity Radiance Field Rendering (CVPR 2025 · similarity 0.88)
- Compact 3D Gaussian Representation for Radiance Field (CVPR 2024 · similarity 0.88)
- Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fields (CVPR 2024 · similarity 0.88)
- pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstruction (CVPR 2024 · similarity 0.88)
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.
