Research map / 3D vision / Gaussian splatting
SRSplat: Feed-Forward Super-Resolution Gaussian Splatting from Sparse Multi-View Images
Xinyuan Hu, Changyue Shi, Chuxiao Yang, Minghao Chen, Jiajun Ding, Tao Wei, Chen Wei, Zhou Yu, Min Tan
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 RGFE reference-guided feature enhancement (ablation) | SRSplat baseline · RealEstate10K (1/4 resolution) | PSNR: 22.20 -> 23.56 | Improved |
| add RGFE reference-guided feature enhancement (ablation) | SRSplat baseline · RealEstate10K (1/4 resolution) | LPIPS: 0.262 -> 0.207 | Improved |
| add TADC texture-aware density control on top of RGFE (ablation) | SRSplat · RealEstate10K (1/4 resolution) | PSNR: 23.56 -> 23.99 | Improved |
| add TADC texture-aware density control on top of RGFE (ablation) | SRSplat · RealEstate10K (1/4 resolution) | LPIPS: 0.207 -> 0.189 | Improved |
| add TADC texture-aware density control on top of RGFE (ablation) | SRSplat · RealEstate10K (1/4 resolution) | SSIM: 0.782 -> 0.802 | Improved |
| use learned texture richness perceptron instead of direct Sobel on LR images (ablation) | SRSplat · RealEstate10K (1/4 resolution) | PSNR: 23.75 -> 23.99 | Improved |
| use learned texture richness perceptron instead of direct Sobel on LR images (ablation) | SRSplat · RealEstate10K (1/4 resolution) | LPIPS: 0.192 -> 0.189 | Improved |
| train on RealEstate10K and test zero-shot on ACID | SRSplat · Cross-dataset RealEstate10K to ACID (zero-shot) | PSNR: +0.85 dB PSNR, +0.075 SSIM, and -0.07 LPIPS | Improved |
| compare feed-forward SRSplat with per-scene optimization SRGS + FSGS | SRSplat · DTU (per-scene comparison) | reconstruction time: 420s -> 0.2s | Improved |
| lower input resolution from 1/2 to 1/8 | SRSplat · RealEstate10K | PSNR: 25.20 -> 21.98 | Hurt |
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.
- pixelSplat — pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstruction (CVPR 2024) · better on 9 of 9 shared results
- MVSplat — MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images (ECCV 2024) · better on 9 of 9 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
- SR3R: Rethinking Super-Resolution 3D Reconstruction With Feed-Forward Gaussian Splatting (CVPR 2026 · similarity 0.92)
- SplatSuRe: Selective Super-Resolution for Multi-view Consistent 3D Gaussian Splatting (CVPR 2026 · similarity 0.91)
- CompactSplat: Spatially Adaptive Gaussian Distribution for Feedforward Scene Reconstruction (NeurIPS 2026 · similarity 0.90)
- S2Gaussian: Sparse-View Super-Resolution 3D Gaussian Splatting (CVPR 2025 · similarity 0.90)
- Learning Efficient Fuse-and-Refine for Feed-Forward 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.89)
- HiSplat: Hierarchical 3D Gaussian Splatting for Generalizable Sparse-View Reconstruction (ICLR 2025 · 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.
