Research map / Point clouds / Point cloud classification
Deep Novel View Synthesis from Colored 3D Point Clouds
Zhenbo Song, Wayne Chen, Dylan Campbell, Hongdong Li
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 RefineNet to the coarse generator (ablation) | Ours-coarse · NYU-V2 scene revealing (4096 points) | PSNR: 14.275 -> 16.439 | Improved |
| add RefineNet to the coarse generator (ablation) | Ours-coarse · NYU-V2 scene revealing (4096 points) | MAE: 0.154 -> 0.117 | Improved |
| add RefineNet to the coarse generator (ablation) | Ours-coarse · NYU-V2 scene revealing (4096 points) | SSIM: 0.414 -> 0.566 | Improved |
| add RefineNet to the coarse generator (ablation) | Ours-coarse · ICL-NUIM novel view synthesis (4096 points) | PSNR: 15.6 -> 18.07 | Improved |
| add RefineNet to the coarse generator (ablation) | Ours-coarse · ICL-NUIM novel view synthesis (4096 points) | SSIM: 0.381 -> 0.579 | Improved |
| increase input points from 4096 to 8192 for refined model (ablation) | Ours-refined · NYU-V2 scene revealing | PSNR: 16.439 -> 16.266 | No clear effect |
| increase input points from 4096 to 8192 for refined model (ablation) | Ours-refined · ICL-NUIM novel view synthesis | PSNR: 18.07 -> 17.75 | No clear effect |
| increase input points from 4096 to 12288 for refined model (ablation) | Ours-refined · NYU-V2 scene revealing | MAE: 0.117 -> 0.125 | Hurt |
| use Ours-refined with 4096 points instead of invsfm using all points | invsfm · NYU-V2 scene revealing | PSNR: 14.507 -> 16.439 | Improved |
| use Ours-refined with 4096 points instead of invsfm using all points | invsfm · ICL-NUIM novel view synthesis | PSNR: 14.737 -> 18.07 | Improved |
| use point-cloud encoder instead of invsfm pipeline for inference speed | invsfm · inference time | seconds: 0.068 -> 0.038 | Improved |
Closest papers by meaning
- Neural Point-Based Graphics (ECCV 2020 · similarity 0.89)
- SynSin: End-to-End View Synthesis From a Single Image (CVPR 2020 · similarity 0.88)
- Point2Pix: Photo-Realistic Point Cloud Rendering via Neural Radiance Fields (CVPR 2023 · similarity 0.88)
- MVPNet: Multi-View Point Regression Networks for 3D Object Reconstruction from A Single Image (AAAI 2019 · similarity 0.88)
- Ponder: Point Cloud Pre-training via Neural Rendering (ICCV 2023 · similarity 0.87)
- View Synthesis with Sculpted Neural Points (ICLR 2023 · similarity 0.87)
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.
