Research map / 3D vision / Gaussian splatting
SA-ResGS: Self-Augmented Residual 3D Gaussian Splatting for Next Best View Selection
Kim Jun-Seong, Tae-Hyun Oh, Eduardo Pérez Pellitero, Youngkyoon Jang
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 physically grounded prefiltering (SA-HashGS, MASt3R) (ablation) | FisherRF · Mip-NeRF 360 | PSNR: 20.642 -> 21.093 | Improved |
| add physically grounded prefiltering (SA-HashGS, MASt3R) (ablation) | FisherRF · Extended dataset | PSNR: 19.490 -> 19.702 | Improved |
| add residual supervision (ResGS) to dynamic selection (ablation) | FisherRF · Mip-NeRF 360 | PSNR: 20.642 -> 20.732 | Improved |
| add residual supervision (ResGS) to dynamic selection (ablation) | FisherRF · Mip-NeRF 360 | LPIPS: 0.450 -> 0.461 | Hurt |
| add residual supervision (ResGS) with fixed view order (ablation) | FisherRF · Mip-NeRF 360 | PSNR: 20.642 -> 21.045 | Improved |
| add residual supervision on top of SA-HashGS (MASt3R) (ablation) | SA-HashGS · Mip-NeRF 360 | PSNR: 21.093 -> 21.410 | Improved |
| add residual supervision on top of SA-HashGS (MASt3R) (ablation) | SA-HashGS · Extended dataset | PSNR: 19.702 -> 20.401 | Improved |
| add residual supervision on top of SA-HashGS (MASt3R) (ablation) | SA-HashGS · Mip-NeRF 360 | LPIPS: 0.441 -> 0.451 | Hurt |
| add residual supervision on top of SA-HashGS (DA-v3) (ablation) | SA-HashGS · Extended dataset | PSNR: 20.025 -> 20.348 | Improved |
| add residual supervision on top of SA-HashGS (DA-v3) (ablation) | SA-HashGS · Extended dataset | LPIPS: 0.357 -> 0.365 | Hurt |
| use 8 instead of 2 extrapolated views in DA-v3 SA-Points (ablation) | SA-ResGS (DA-v3) · SA-Point backbone ablation | PSNR: 21.247 -> 21.429 | Improved |
| add full loss to subset-only residual loss (ablation) | ResGS · Mip-NeRF 360 | qualitative | 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.
- CrossScore — CrossScore: A Multi-View Approach to Image Evaluation and Scoring (ECCV 2024) · better on 3 of 3 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
- D$^2$GS: Depth-and-Density Guided Gaussian Splatting for Stable and Accurate Sparse-View Reconstruction (ICLR 2026 · similarity 0.90)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.90)
- MAC-Splat: Multi-Attribute Consistency for High-Fidelity Sparse-View Reconstruction (ECCV 2026 · similarity 0.90)
- SparseSplat: Towards Applicable Feed-Forward 3D Gaussian Splatting with Pixel-Unaligned Prediction (CVPR 2026 · similarity 0.89)
- Quantifying and Alleviating Co-Adaptation in Sparse-View 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.88)
- Metropolis-Hastings Sampling for 3D Gaussian Reconstruction (NeurIPS 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.
