Research map / 3D vision / Gaussian splatting
SupIR-GS: Thermal Infrared Super-Resolution Novel View Synthesis with Imaging-Calibrated 3D Gaussian Splatting
Jin Liu, Haodong Li, Jiagang Chen, Dabin leng, Jiguang Li, Zhao Huang, Xiaoshuai Zhang, Qi Xu, Zhiwen Zheng, Xingru Huang
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 PhysIR-Cal alone (ablation) | SRGS-like baseline · TI-NSD ablation Indoor | PSNR (dB): 32.086 -> 33.692 | Improved |
| add IRTone-Adapter alone (ablation) | SRGS-like baseline · TI-NSD ablation Indoor | PSNR (dB): 32.086 -> 31.700 | Hurt |
| add Frequency-Aware curriculum alone (ablation) | SRGS-like baseline · TI-NSD ablation Outdoor | PSNR (dB): 27.381 -> 27.263 | Hurt |
| add Frequency-Aware curriculum after PhysIR-Cal (ablation) | PhysIR-Cal · TI-NSD ablation Outdoor | PSNR (dB): 28.449 -> 29.800 | Improved |
| add IRTone-Adapter after PhysIR-Cal + Frequency-Aware (ablation) | PhysIR-Cal + Frequency-Aware · TI-NSD ablation Indoor | PSNR (dB): 34.276 -> 34.806 | Improved |
| add IRTone-Adapter after PhysIR-Cal + Frequency-Aware (ablation) | PhysIR-Cal + Frequency-Aware · TI-NSD ablation Outdoor | PSNR (dB): 29.800 -> 29.935 | No clear effect |
| use SupIR-GS instead of SRGS | SRGS · TI-NSD (4x SR from LR) Indoor | PSNR (dB): 32.09 -> 34.81 | Improved |
| use SupIR-GS instead of SRGS | SRGS · ThermoScenes (4x SR from LR) Outdoor | PSNR (dB): 30.76 -> 30.40 | Hurt |
| use SupIR-GS instead of Bicubic+3DGS | Bicubic + 3DGS · ThermoScenes (4x SR from LR) Indoor | MAEroi: 1.6318 -> 1.7903 | Hurt |
| use SupIR-GS instead of 3DGS (4x) | 3DGS (4x) · TI-NSD (4x SR from LR) Outdoor | PSNR (dB): 21.86 -> 29.94 | Improved |
| use SupIR-GS instead of 3DGS | 3DGS · Efficiency | training time (min): 10.39 -> 16.42 | Hurt |
| use SupIR-GS instead of SRGS | SRGS · Efficiency | FPS: 313.75 -> 381.10 | Improved |
Closest papers by meaning
- PhysIR-Splat: Physically Consistent Thermal Infrared Radiative Transfer in 3D Gaussian Splatting (CVPR 2026 · similarity 0.92)
- Thermal3D-GS: Physics-induced 3D Gaussians for Thermal Infrared Novel-view Synthesis (ECCV 2024 · similarity 0.90)
- ThermalGaussian: Thermal 3D Gaussian Splatting (ICLR 2025 · similarity 0.87)
- GaussHDR: High Dynamic Range Gaussian Splatting via Learning Unified 3D and 2D Local Tone Mapping (CVPR 2025 · similarity 0.86)
- NTR-Gaussian: Nighttime Dynamic Thermal Reconstruction with 4D Gaussian Splatting Based on Thermodynamics (CVPR 2025 · similarity 0.86)
- S2Gaussian: Sparse-View Super-Resolution 3D Gaussian Splatting (CVPR 2025 · similarity 0.86)
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.
