Research map / 3D vision / Gaussian splatting
Perceptual Quality Assessment of 3D Gaussian Splatting: A Subjective Dataset and Prediction Metric
Zhaolin Wan, Yining Diao, Jingqi Xu, Hao Wang, Zhiyang Li, Xiaopeng Fan, Wangmeng Zuo, Debin Zhao
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 GAT module (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | PLCC: 0.7914 -> 0.8213 | Improved |
| add GAT module (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | SRCC: 0.7496 -> 0.7738 | Improved |
| add attention pooling instead of global pooling (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | SRCC: 0.7359 -> 0.7738 | Improved |
| add monotonicity loss L_mon (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | SRCC: 0.7124 -> 0.7738 | Improved |
| add linearity loss L_lin (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | PLCC: 0.7161 -> 0.8213 | Improved |
| add linearity loss L_lin (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | RMSE: 1.5948 -> 1.0798 | Improved |
| change loss weights to lambda1=0.4, lambda2=0.6 (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | PLCC: 0.8213 -> 0.8244 | No clear effect |
| shift loss weight toward L_lin (lambda1=0.8, lambda2=0.2) (ablation) | GSOQA · 3DGS-QA (5-fold cross-validation) | PLCC: 0.8213 -> 0.7953 | Hurt |
| operate directly on native Gaussian primitives instead of rendered images or point clouds | GSOQA vs IQA/VQA/3DQA baselines · 3DGS-QA (5-fold cross-validation) | PLCC: 0.5698 -> 0.8213 | Improved |
Closest papers by meaning
- IBGS: Image-Based Gaussian Splatting (NeurIPS 2025 · similarity 0.92)
- 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors (NeurIPS 2024 · similarity 0.91)
- Quantifying and Alleviating Co-Adaptation in Sparse-View 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.91)
- GaussianPro: 3D Gaussian Splatting with Progressive Propagation (ICML 2024 · similarity 0.91)
- LP-3DGS: Learning to Prune 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- Perceptual-GS: Scene-adaptive Perceptual Densification for Gaussian Splatting (ICML 2025 · similarity 0.91)
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.
