Efficient LiDAR Reflectance Compression via Scanning Serialization

2025-05-14

AI Summary

To address the low compression efficiency and weak modeling capability of neural methods for LiDAR point cloud reflectance attributes, this paper proposes SerLiC, a serialized compression framework. SerLiC is the first to map 3D point clouds into 1D sequences according to scan order, constructing contextual representations by fusing scan indices, radial distances, and prior reflectance. It introduces a device-centric modeling perspective and designs a dual-parallel Mamba state-space architecture to enable efficient autoregressive dependency modeling with minimal parameters. Experiments demonstrate that SerLiC achieves over 2× compression ratio, reduces bit-rate by 22% compared to SOTA methods, and requires only 2% of their parameter count. Its lightweight variant operates at 10.3 fps with merely 111K parameters, striking an exceptional balance among compression performance, real-time inference, and deployment efficiency.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Reflectance attributes in LiDAR point clouds provide essential information for downstream tasks but remain underexplored in neural compression methods. To address this, we introduce SerLiC, a serialization-based neural compression framework to fully exploit the intrinsic characteristics of LiDAR reflectance. SerLiC first transforms 3D LiDAR point clouds into 1D sequences via scan-order serialization, offering a device-centric perspective for reflectance analysis. Each point is then tokenized into a contextual representation comprising its sensor scanning index, radial distance, and prior reflectance, for effective dependencies exploration. For efficient sequential modeling, Mamba is incorporated with a dual parallelization scheme, enabling simultaneous autoregressive dependency capture and fast processing. Extensive experiments demonstrate that SerLiC attains over 2x volume reduction against the original reflectance data, outperforming the state-of-the-art method by up to 22% reduction of compressed bits while using only 2% of its parameters. Moreover, a lightweight version of SerLiC achieves>10 fps (frames per second) with just 111K parameters, which is attractive for real-world applications.

Problem

• Compressing LiDAR reflectance efficiently using neural methods
• Transforming 3D point clouds into 1D sequences for analysis
• Achieving high compression rates with minimal parameters

Innovation

• Scan-order serialization for 1D sequence transformation
• Tokenization with contextual sensor and reflectance data
• Dual parallelization Mamba for efficient sequential modeling