Research map / Point clouds / Point cloud classification
PFGS: High Fidelity Point Cloud Rendering via Feature Splatting
Jiaxu Wang, Zhang Ziyi, Junhao He, Renjing Xu
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 two-stage training (pretrain Gaussian regressor before joint training) (ablation) | PFGS · ScanNet | PSNR: 15.92 -> 19.86 | Improved |
| add multiscale decoder (vs single-scale UNet) (ablation) | PFGS · ScanNet | PSNR: 18.27 -> 19.86 | Improved |
| add progressive loss terms (ablation) | PFGS · ScanNet | PSNR: 19.19 -> 19.86 | Improved |
| add multiscale loss terms (ablation) | PFGS · ScanNet | PSNR: 19.37 -> 19.86 | Improved |
| add frequency reconstruction loss L_mfr (ablation) | PFGS · ScanNet | LPIPS: 0.478 -> 0.452 | Improved |
| run decoder recurrently twice instead of once (ablation) | PFGS · ScanNet | PSNR: 19.19 -> 19.86 | Improved |
| increase decoder loops from 2 to 3 (ablation) | PFGS · ScanNet | PSNR: 19.86 -> 19.77 | No clear effect |
| train on dense points, test on sparse points (ablation) | PFGS · DTU (train 1, test 0.2) | PSNR: 12.78 vs 25.88 at matched density | Hurt |
| increase test point count with model trained on 120k points (ablation) | PFGS · THuman2.0 | PSNR: 31.60 -> 39.18 | Improved |
| replace Trivol (NeRF-based) with PFGS under added point noise | point cloud renderer · THuman2.0 noise 3e-3 | PSNR: 20.04 -> 30.85 | Improved |
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.
- Trivol — TriVol: Point Cloud Rendering via Triple Volumes (CVPR 2023) · better on 6 of 6 shared results
- NPBG++ — NPBG++: Accelerating Neural Point-Based Graphics (CVPR 2022) · better on 3 of 4 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
- Sparse Point Cloud Patches Rendering via Splitting 2D Gaussians (CVPR 2025 · similarity 0.90)
- SplatFormer: Point Transformer for Robust 3D Gaussian Splatting (ICLR 2025 · similarity 0.90)
- Compact 3D Gaussian Representation for Radiance Field (CVPR 2024 · similarity 0.89)
- LP-3DGS: Learning to Prune 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.89)
- GaussianPro: 3D Gaussian Splatting with Progressive Propagation (ICML 2024 · similarity 0.89)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 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.
