Research map / Point clouds / Point cloud classification
Efficient Hierarchical Entropy Model for Learned Point Cloud Compression
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 EHEM instead of OctAttention | Octree entropy modeling on SemanticKITTI · SemanticKITTI, D=16 | decoding time: 708 -> 3.01 | Helped |
| Use grouped context instead of auto-regressive context (ablation) | EHEM entropy model · SemanticKITTI | decoding latency | Helped |
| Use grouped context instead of auto-regressive context (ablation) | EHEM entropy model · SemanticKITTI | compression performance | No clear effect |
| Use hierarchical attention instead of global self-attention (ablation) | EHEM entropy model · SemanticKITTI, N=1024 | D1 PSNR | Helped |
| Extend context length from N=1024 to N=8192 (ablation) | EHEM entropy model · SemanticKITTI | rate-distortion performance | Helped |
| Use a deeper self-attention network (ablation) | EHEM entropy model · SemanticKITTI | rate-distortion performance | Helped |
| Reduce self-attention layers and channel dimension as in Light EHEM (ablation) | EHEM entropy model · SemanticKITTI | FLOPs: 184.4G -> 102.9G | Helped |
| Reduce self-attention layers and channel dimension as in Light EHEM (ablation) | EHEM entropy model · SemanticKITTI | rate-distortion performance | Hurt |
| Use EHEM instead of OctAttention | Point-cloud compression · SemanticKITTI | average bitrate: 19.47% bitrate reduction | Helped |
| Use EHEM instead of G-PCC | Point-cloud compression · SemanticKITTI | average bitrate: 28.89% | Helped |
| Use EHEM instead of OctAttention | Octree entropy modeling · SemanticKITTI, D=16 | encoding time: 0.66 -> 2.53 | Hurt |
Closest papers by meaning
- Generalized Gaussian Entropy Model for Point Cloud Attribute Compression with Dynamic Likelihood Intervals (CVPR 2025 · similarity 0.86)
- Learned Image Compression with Dictionary-based Entropy Model (CVPR 2025 · similarity 0.86)
- TopNet: Transformer-Efficient Occupancy Prediction Network for Octree-Structured Point Cloud Geometry Compression (CVPR 2025 · similarity 0.86)
- Diversify, Contextualize, and Adapt: Efficient Entropy Modeling for Neural Image Codec (NeurIPS 2024 · similarity 0.85)
- UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach (AAAI 2025 · similarity 0.85)
- Laplacian-guided Entropy Model in Neural Codec with Blur-dissipated Synthesis (CVPR 2024 · similarity 0.84)
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.