Research map / Point clouds / Point cloud classification
Inferring Point Cloud Quality via Graph Similarity
Qi Yang, Zhan Ma, Yiling Xu, Zhu Li, Jun Sun
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 graph-similarity GraphSIM instead of point-wise distance metrics | M-p2po · SJTU-PCQA (People) | SROCC: 0.79 -> 0.88 | Improved |
| use GraphSIM instead of PSNR_YUV on geometric distortions | PSNR_YUV · SJTU-PCQA (People) | PLCC: DS 0.74 -> 0.97, GGN 0.86 -> 0.97 | Improved |
| use GraphSIM instead of PSNR_YUV on color noise | PSNR_YUV · SJTU-PCQA (CN impairment) | PLCC: 0.97 -> 0.90 | Hurt |
| use GraphSIM instead of PSNR_YUV | PSNR_YUV · IRPC (all samples) | SROCC: 0.50 -> 0.76 | Improved |
| use GraphSIM instead of M-p2po | M-p2po · IRPC (all samples) | PLCC: 0.71 -> 0.94 | Improved |
| swap color space among RGB, YUV, GCM (ablation) | GraphSIM · SJTU-PCQA (People) | PLCC: 0.8948 vs 0.8941 vs 0.8896 | No clear effect |
| use high-pass resampling instead of random resampling (ablation) | GraphSIM · SJTU-PCQA (People) | PLCC: 0.8836 -> 0.8896 | Improved |
| increase neighbor dimension theta/B (ablation) | GraphSIM · SJTU-PCQA (People) | performance | Improved |
| increase k nearest neighbors for graph scale (ablation) | GraphSIM · SJTU-PCQA (People) | performance | Improved |
| use averaging instead of multiplication for pooling features and channels (C1 vs C4) (ablation) | GraphSIM · SJTU-PCQA (People) | PLCC: 0.7827 -> 0.8966 | Improved |
| pool features by multiplication and channels by average (C2) vs average both (C1) (ablation) | GraphSIM · SJTU-PCQA (People) | PLCC: 0.8966 -> 0.8896 | No clear effect |
| add color to coordinate signal (M1) instead of coordinate only (G1) (ablation) | GraphSIM · SJTU-PCQA (People) | PLCC: 0.8822 -> 0.9381 | Improved |
Closest papers by meaning
- No-Reference Point Cloud Quality Assessment via Domain Adaptation (CVPR 2022 · similarity 0.87)
- Points Meet Pixels: Bridging 2D Vision-Language Model and 3D Perception Gaps for Point Cloud Quality Assessment (AAAI 2026 · similarity 0.87)
- CLIP-PCQA: Exploring Subjective-Aligned Vision-Language Modeling for Point Cloud Quality Assessment (AAAI 2025 · similarity 0.87)
- GraphFit: Learning Multi-Scale Graph-Convolutional Representation for Point Cloud Normal Estimation (ECCV 2022 · similarity 0.87)
- Learning Disentangled Representations for Perceptual Point Cloud Quality Assessment via Mutual Information Minimization (NeurIPS 2024 · similarity 0.86)
- QD-PCQA: Quality-Aware Domain Adaptation for Point Cloud Quality Assessment (CVPR 2026 · 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.
