Research map / Point clouds / Point cloud classification
TriVol: Point Cloud Rendering via Triple Volumes
Tao Hu, Xiaogang Xu, Ruihang Chu, Jiaya Jia
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 |
|---|---|---|---|
| Use TriVol instead of the strongest reported baseline | Point cloud rendering · ScanNet | PSNR: 17.65 -> 18.56 | Improved |
| Use TriVol instead of the strongest reported baseline | Point cloud rendering · ScanNet | SSIM: 0.694 -> 0.734 | Improved |
| Use TriVol instead of the strongest reported baseline | Point cloud rendering · ScanNet | LPIPS: 0.517 -> 0.473 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · ShapeNet Car | PSNR: 25.73 -> 27.22 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · ShapeNet Car | SSIM: 0.897 -> 0.927 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · ShapeNet Car | LPIPS: 0.107 -> 0.084 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · Google Scanned Objects Shoe | PSNR: 29.84 -> 31.24 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · Google Scanned Objects Shoe | SSIM: 0.938 -> 0.961 | Improved |
| Use TriVol instead of Point-NeRF | Point cloud rendering · Google Scanned Objects Shoe | LPIPS: 0.069 -> 0.045 | Improved |
| Use the grouping encoder rather than PointNet, PointNet++, or Sparse ConvNet (ablation) | TriVol encoder · ScanNet | PSNR: 18.03 -> 18.56; 18.26 -> 18.56; 18.50 -> 18.56 | Improved |
| Use TriVol with G=16, S=256^2 rather than dense voxels at 128^3 (ablation) | TriVol representation · ShapeNet Car | PSNR: 25.53 -> 27.22 | Improved |
| Increase TriVol resolution from S=128^2 to S=256^2 at G=16 (ablation) | TriVol representation · ShapeNet Car | PSNR: 24.84 -> 27.22 | Improved |
Later papers that compared with this method
2 later accepted papers report TriVol as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Sparse Point Cloud Patches Rendering via Splitting 2D Gaussians (CVPR 2025 · better on 6 of 6 shared results)
- PFGS: High Fidelity Point Cloud Rendering via Feature Splatting (ECCV 2024 · better on 6 of 6 shared results)
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.
- Point-NeRF — Point-NeRF: Point-Based Neural Radiance Fields (CVPR 2022) · better on 7 of 7 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
- Ponder: Point Cloud Pre-training via Neural Rendering (ICCV 2023 · similarity 0.88)
- Point2Real: Bridging the Gap between Point Cloud and Realistic Image for Open-World 3D Recognition (AAAI 2024 · similarity 0.87)
- Learning Efficient Point Cloud Generation for Dense 3D Object Reconstruction (AAAI 2018 · similarity 0.87)
- 3DMambaIPF: A State Space Model for Iterative Point Cloud Filtering via Differentiable Rendering (AAAI 2025 · similarity 0.87)
- UniPre3D: Unified Pre-training of 3D Point Cloud Models with Cross-Modal Gaussian Splatting (CVPR 2025 · similarity 0.86)
- Duoduo CLIP: Efficient 3D Understanding with Multi-View Images (ICLR 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.
