UniPCGC: Towards Practical Point Cloud Geometry Compression via an Efficient Unified Approach

TL;DR

UniPCGC introduces a practical, unified point cloud geometry compression framework that jointly supports lossless, lossy, variable rate, and variable complexity. It combines Uneven 8-Stage Lossless Coding (UELC) for lossless coordinate coding with a Variable Rate and Complexity Module (VRCM) for lossy feature coding, enabling dynamic switching and joint optimization. Empirical results show a lossless CR improvement of $8.1\%$ and a lossy BD-Rate improvement of $14.02\%$ over strong baselines, while maintaining real-world feasibility and flexible rate control. The approach advances practical pcg compression by reducing decoding complexity and enabling adaptable performance across bandwidth requirements.

Abstract

Learning-based point cloud compression methods have made significant progress in terms of performance. However, these methods still encounter challenges including high complexity, limited compression modes, and a lack of support for variable rate, which restrict the practical application of these methods. In order to promote the development of practical point cloud compression, we propose an efficient unified point cloud geometry compression framework, dubbed as UniPCGC. It is a lightweight framework that supports lossy compression, lossless compression, variable rate and variable complexity. First, we introduce the Uneven 8-Stage Lossless Coder (UELC) in the lossless mode, which allocates more computational complexity to groups with higher coding difficulty, and merges groups with lower coding difficulty. Second, Variable Rate and Complexity Module (VRCM) is achieved in the lossy mode through joint adoption of a rate modulation module and dynamic sparse convolution. Finally, through the dynamic combination of UELC and VRCM, we achieve lossy compression, lossless compression, variable rate and complexity within a unified framework. Compared to the previous state-of-the-art method, our method achieves a compression ratio (CR) gain of 8.1\% on lossless compression, and a Bjontegaard Delta Rate (BD-Rate) gain of 14.02\% on lossy compression, while also supporting variable rate and variable complexity.

Figures (5)

• Figure 1: Performance comparisons for the proposed method and the other methods. The left figure shows the CR gain obtained by the proposed UniPCGC method in lossless compression. The right figure shows the BD-rate gain obtained by the proposed UniPCGC method in lossy compression.
• Figure 2: llustration of the proposed UniPCGC framework. It mainly consists of two parts: coordinate coding and feature coding. Coordinate coding is performed using Uneven 8-Stage Lossless Code (UELC) at each scale. Feature coding is performed using Variable Rate and Complexity Module (VRCM), which mainly includes encoder, modulation, demodulation, decoder, complexity and bit allocation and Factorized Entropy Model. AE/AD stands for arithmetic encoding and arithmetic decoding.
• Figure 3: Illustration of the proposed Uneven 8-Stage Lossless Coder (UELC). In each stage, the previously coding groups and their features are regarded as prior information to better estimate the occupancy probability of groups in the current stage.
• Figure 4: Detailed architecture of proposed VRCM. It also shows the architecture of channel level bit allocation module, One-Stage Lossless Coder (OLC / DOLC) and Dynamic Feature Extraction Layer (DFEL).
• Figure 5: Performance comparison using rate-distortion curves.