Research map / Point clouds / Point cloud classification
Towards Practical Lossless Neural Compression for LiDAR Point Clouds
pengpeng yu, Haoran Li, Runqing Jiang, Dingquan Li, Jing Wang, Liang Lin, Yulan Guo
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 GRED to the baseline octree codec (ablation) | Baseline · KITTI D1 | BD-Rate vs Baseline (%): 0.00 -> -3.78 | Improved |
| add GRED to the baseline octree codec (ablation) | Baseline · KITTI D1 | BD-PSNR vs Baseline (dB): 0.00 -> 0.40 | Improved |
| add XFP on top of GRED (ablation) | Baseline + GRED · KITTI D1 | BD-Rate vs Baseline (%): -3.78 -> -13.22 | Improved |
| add XFP on top of GRED (ablation) | Baseline + GRED · KITTI D1 | BD-PSNR vs Baseline (dB): 0.40 -> 1.46 | Improved |
| add XFP on top of GRED (ablation) | Baseline + GRED · KITTI D1 | BD-Rate vs G-PCC octree (%): -13.44 -> -21.93 | Improved |
| add GRED to the baseline octree codec (ablation) | Baseline · KITTI D1 | BD-Rate vs G-PCC octree (%): -10.04 -> -13.44 | Improved |
| add GRED to the baseline octree codec (ablation) | Baseline · KITTI D1 | BD-PSNR vs G-PCC octree (dB): 1.13 -> 1.49 | Improved |
| use integer-only inference instead of floating point | Ours · KITTI D1 | BD-Rate vs RENO (%): -15.582 -> -14.289 | Mixed |
| use integer-only inference instead of floating point | Ours · Ford D1 | BD-Rate vs RENO (%): -11.049 -> -6.875 | Hurt |
| use integer-only inference instead of floating point | Ours · KITTI encoding time (RTX 4090) | encoding time (s): 0.075s -> 0.060s | Improved |
| run integer model on RTX 5880 instead of RTX 4090 | Ours (Integer) · KITTI encoding time | encoding time (s): 0.060s -> 0.047s | Improved |
| run integer model on Tesla V100 instead of RTX 4090 | Ours (Integer) · KITTI encoding time | encoding time (s): 0.060s -> 0.166s | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- RENO — RENO: Real-Time Neural Compression for 3D LiDAR Point Clouds (CVPR 2025) · better on 0 of 1 shared result
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
- Low-Latency Neural LiDAR Compression with 2D Context Models (ICLR 2026 · similarity 0.92)
- RENO: Real-Time Neural Compression for 3D LiDAR Point Clouds (CVPR 2025 · similarity 0.90)
- msLPCC: A Multimodal-Driven Scalable Framework for Deep LiDAR Point Cloud Compression (AAAI 2024 · similarity 0.89)
- Efficient LiDAR Reflectance Compression via Scanning Serialization (ICML 2025 · similarity 0.88)
- UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach (AAAI 2025 · similarity 0.88)
- SCP: Spherical-Coordinate-Based Learned Point Cloud Compression (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.
