Research map / Point clouds / Point cloud classification
RIDDLE: Lidar Data Compression With Range Image Deep Delta Encoding
Xuanyu Zhou, Charles R. Qi, Yin Zhou, Dragomir Anguelov
Check 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 |
|---|---|---|---|
| compress range images with deep delta prediction | RIDDLE on Waymo Open Dataset · Waymo Open Dataset validation, Chamfer distance around 0.005 | bitrate: 10.78 bpp to 3.65 bpp; more than 65% | Improved |
| compress range images with RIDDLE | Waymo Open Dataset point clouds · Waymo Open Dataset | Chamfer distance: LASzip: 0.0048 at 10.62 bpp; RIDDLE: 0.0026 at 4.51 bpp | Improved |
| use an adapted PointNet predictor instead of linear or CNN predictors (ablation) | RIDDLE · range prediction, quantization 0.1m | prediction accuracy: 64.62 -> 65.75 | Improved |
| add anchor classification and regression instead of direct regression or multi-bin prediction (ablation) | RIDDLE · range prediction, quantization 0.1m | prediction accuracy: 61.64 -> 65.75 | Improved |
| add prior-scan image-patch context (ablation) | RIDDLE-T · range prediction, quantization 0.1m | prediction accuracy: 65.75 -> 67.34 | Improved |
| use pose-aware 100-point temporal neighbors instead of same-row-and-column prior-scan pixels (ablation) | RIDDLE-T · range prediction, quantization 0.1m | prediction accuracy: 67.34 -> 69.23 | Improved |
| select sparse representation with arithmetic encoding for sparse residuals (ablation) | RIDDLE · residual compression, quantization 0.1m | bitrate: 2.28 vs 2.33 bpp | Improved |
| select varints with LZMA for less sparse residuals (ablation) | RIDDLE · residual compression, quantization 0.02m | bitrate: 4.17 -> 4.08 bpp | Improved |
| enlarge the predictor context from 5 × 10 to 10 × 10 (ablation) | RIDDLE · intra-frame range prediction, quantization 0.1m | prediction accuracy: 65.21 -> 65.75 | Improved |
| include spherical azimuth and inclination rather than relative row/column indices (ablation) | RIDDLE · intra-frame range prediction, 5 × 10 context | prediction accuracy: 65.02 -> 65.21 | Improved |
| quantize range, intensity, and elongation at the reported precisions (ablation) | Waymo Open Dataset PointPillars detection · vehicle and pedestrian detection | mAP: 69.59 vs 69.59; 65.62 vs 65.62 | No clear effect |
Later papers that compared with this method
2 later accepted papers report RIDDLE as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Low-Latency Neural LiDAR Compression with 2D Context Models (ICLR 2026 · better on 3 of 3 shared results)
- R²D-LPCC: Relevance-Ranking Guided Region-Adaptive Dynamic LiDAR Point Cloud Compression (AAAI 2026 · better on 1 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
- Point Cloud Compression with Range Image-Based Entropy Model for Autonomous Driving (ECCV 2022 · similarity 0.91)
- BitRIC: Efficient Neural Compression of LiDAR Range Images via Hierarchical Bitplanes (ECCV 2026 · similarity 0.88)
- RENO: Real-Time Neural Compression for 3D LiDAR Point Clouds (CVPR 2025 · similarity 0.85)
- msLPCC: A Multimodal-Driven Scalable Framework for Deep LiDAR Point Cloud Compression (AAAI 2024 · similarity 0.84)
- SCP: Spherical-Coordinate-Based Learned Point Cloud Compression (AAAI 2024 · similarity 0.84)
- Towards Realistic Scene Generation with LiDAR Diffusion Models (CVPR 2024 · similarity 0.83)
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.
