Research map / 3D vision / Gaussian splatting
PhysIR-Splat: Physically Consistent Thermal Infrared Radiative Transfer in 3D Gaussian Splatting
Jingyuan Gao, Yumeng Hu, Fei Gao, Mingjin Zhang
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 |
|---|---|---|---|
| initialize with VGGT-IR instead of COLMAP (ablation) | PhysIR-Splat · TI-NSD average | PSNR: 36.47 -> 37.33 | Improved |
| add IR-specific opacity instead of shared opacity (ablation) | PhysIR-Splat · TI-NSD average | PSNR: 36.46 -> 37.33 | Improved |
| add distance-dependent atmospheric transmittance (ablation) | PhysIR-Splat · TI-NSD average | PSNR: 36.56 -> 37.33 | Improved |
| use VGGT-IR instead of VGGT for pose estimation on LWIR | feed-forward initializer · Multi-Spectral Dataset desk1-xyz-dim2 | APE RMSE translation: 0.2133 -> 0.0823 | Improved |
| use VGGT-IR instead of COLMAP for pose estimation on LWIR | pose estimation · Multi-Spectral Dataset desk2-halfsphere-ic | APE RMSE translation: 0.0643 -> 0.0176 | Improved |
| embed TIR radiative transfer in the 3DGS renderer (physical attributes per Gaussian) | 3DGS · RGBT-Scenes THERMAL average | PSNR: 25.73 -> 27.15 (vs ThermalGaussian, COLMAP init) | Improved |
| embed TIR radiative transfer in the 3DGS renderer with COLMAP init | 3DGS · TI-NSD average | PSNR: 35.97 -> 36.47 (vs Veta-GS) | Improved |
| physical modeling effect on specular materials (glass, polished steel) | emissivity estimation · RGBT-Scenes emissivity analysis | emissivity agreement | Hurt |
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.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 2 of 2 shared results
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
- SupIR-GS: Thermal Infrared Super-Resolution Novel View Synthesis with Imaging-Calibrated 3D Gaussian Splatting (ECCV 2026 · similarity 0.92)
- ThermalGaussian: Thermal 3D Gaussian Splatting (ICLR 2025 · similarity 0.90)
- NTR-Gaussian: Nighttime Dynamic Thermal Reconstruction with 4D Gaussian Splatting Based on Thermodynamics (CVPR 2025 · similarity 0.89)
- 3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes (CVPR 2025 · similarity 0.88)
- pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstruction (CVPR 2024 · similarity 0.87)
- SplatFormer: Point Transformer for Robust 3D Gaussian Splatting (ICLR 2025 · similarity 0.87)
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.
