ELiC: Efficient LiDAR Geometry Compression via Cross-Bit-depth Feature Propagation and Bag-of-Encoders

2025-11-17

AI Summary

Existing hierarchical LiDAR geometric compression methods model different bit depths independently and repeatedly estimate local context layer-by-layer, resulting in high redundancy and suboptimal entropy coding efficiency. To address this, we propose a real-time, efficient compression framework: (1) cross-bit-depth feature propagation, where low-bit-depth features are reused to guide high-bit-depth prediction; (2) a bag-of-encoders dynamic selection mechanism that adapts to varying semantic complexity across hierarchy levels; and (3) a Morton-order-preserving hierarchical voxelization structure, eliminating per-layer reordering and enhancing contextual modeling consistency and entropy coding performance. Integrated with a context-aware entropy model, our method achieves state-of-the-art compression ratios on the Ford and SemanticKITTI datasets, while attaining real-time encoding throughput—significantly outperforming existing approaches.

Technology Category

• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Learning on the Edge & Model Compression
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Hierarchical LiDAR geometry compression encodes voxel occupancies from low to high bit-depths, yet prior methods treat each depth independently and re-estimate local context from coordinates at every level, limiting compression efficiency. We present ELiC, a real-time framework that combines cross-bit-depth feature propagation, a Bag-of-Encoders (BoE) selection scheme, and a Morton-order-preserving hierarchy. Cross-bit-depth propagation reuses features extracted at denser, lower depths to support prediction at sparser, higher depths. BoE selects, per depth, the most suitable coding network from a small pool, adapting capacity to observed occupancy statistics without training a separate model for each level. The Morton hierarchy maintains global Z-order across depth transitions, eliminating per-level sorting and reducing latency. Together these components improve entropy modeling and computation efficiency, yielding state-of-the-art compression at real-time throughput on Ford and SemanticKITTI. Code and models will be released upon publication.

Problem

• Improves LiDAR geometry compression efficiency by reusing features across bit depths
• Adapts encoding capacity per depth level without training separate models
• Maintains global ordering across depth transitions to reduce computational latency

Innovation

• Cross-bit-depth feature propagation reuses lower depth features
• Bag-of-Encoders adaptively selects optimal coding networks
• Morton-order-preserving hierarchy eliminates per-level sorting