Research map / Point clouds / Point cloud classification
AnyPcc: Compressing Any Point Cloud with a Single Universal Model
Kangli Wang, Qianxi Yi, Yuqi Ye, Shihao Li, Wei Gao
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 spatial grouping (ablation) | UCM · AnyPcc benchmark (lossless) | CR-Gain vs baseline: 0.00% -> -6.56% | Improved |
| add channel grouping alone (ablation) | UCM · AnyPcc benchmark (lossless) | CR-Gain vs baseline: 0.00% -> 0.13% | Hurt |
| add channel grouping on top of SC+SG (ablation) | UCM · AnyPcc benchmark (lossless) | CR-Gain vs baseline: -7.74% -> -9.88% | Improved |
| add spatial convolution on top of SG+CG (ablation) | UCM · AnyPcc benchmark (lossless) | CR-Gain vs baseline: -6.50% -> -9.88% | Improved |
| add spatial grouping on top of SC+CG (ablation) | UCM · AnyPcc benchmark (lossless) | CR-Gain vs baseline: -5.33% -> -9.88% | Improved |
| add IAFT (800 iterations) (ablation) | UCM · GS (3DGS, OOD), 64 channels | bits per point: 5.32 -> 5.04 | Improved |
| add IAFT (ablation) | UCM · GS (3DGS, OOD) | entropy-coded bpp: 13.307 -> 11.424 | Improved |
| increase channels from 64 to 128 (ablation) | UCM · GS (3DGS, OOD) | bits per point: 5.32 -> 5.25 | Mixed |
| use unified model (AnyPcc-U) instead of per-category model on OOD | AnyPcc · OOD datasets | bits per point: 7.30 -> 7.06 (VGGT); 10.93 -> 10.79 (S3DIS); 11.65 -> 11.74 (GS) | Mixed |
| AnyPcc vs GPCC v23 | AnyPcc · AnyPcc benchmark (lossless) | CR-Gain over GPCC: -11.93% | Improved |
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.
- RENO — RENO: Real-Time Neural Compression for 3D LiDAR Point Clouds (CVPR 2025) · better on 5 of 5 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
- UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach (AAAI 2025 · similarity 0.90)
- A Versatile Point Cloud Compressor Using Universal Multiscale Conditional Coding – Part I: Geometry (TPAMI 2024 · similarity 0.89)
- Deep Learning-Based Point Cloud Compression: An In-Depth Survey and Benchmark (TPAMI 2025 · similarity 0.89)
- Fast Point Cloud Geometry Compression with Context-based Residual Coding and INR-based Refinement (ECCV 2024 · similarity 0.88)
- TopNet: Transformer-Efficient Occupancy Prediction Network for Octree-Structured Point Cloud Geometry Compression (CVPR 2025 · similarity 0.88)
- patchDPCC: A Patchwise Deep Compression Framework for Dynamic Point Clouds (AAAI 2024 · 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.
