Geometry-based point cloud compression (G-PCC), an international standard designed by MPEG, provides a generic framework for compressing diverse types of point clouds while ensuring interoperability across applications and devices. However, G-PCC underperforms compared to recent deep learning-based PCC methods despite its lower computational power consumption. To enhance the efficiency of G-PCC without sacrificing its interoperability or computational flexibility, we propose the first compression-oriented point cloud voxelization network jointly optimized with a differentiable G-PCC surrogate model. The surrogate model mimics the rate-distortion behavior of the non-differentiable G-PCC codec, enabling end-to-end gradient propagation. The versatile voxelization network adaptively transforms input point clouds using learning-based voxelization and effectively manipulates point clouds via global scaling, fine-grained pruning, and point-level editing for rate-distortion trade-off. During inference, only the lightweight voxelization network is prepended to the G-PCC encoder, requiring no modifications to the decoder, thus introducing no computational overhead for end users. Extensive experiments demonstrate a 38.84% average BD-rate reduction over G-PCC. By bridging classical codecs with deep learning, this work offers a practical pathway to enhance legacy compression standards while preserving their backward compatibility, making it ideal for real-world deployment.

Dynamic point cloud compression is essential for efficient 3D visual data transmission and storage. To achieve satisfactory coding efficiency, existing learning-based frameworks typically design separate networks for intra-frame and inter frame geometry compression, resulting in frequent model switching, buffering, and loading, which increases bus bandwidth consumption and latency within hardware systems. Our experiments show that model switching can account for 30.4% of per-frame coding time, especially in longer sequences with fewer frames per GOF (Group of Frames). To overcome this, this paper proposes a unified geometry compression framework that handles both intra-frame and inter-frame coding within a single network. Central to our approach is a 4D sparse tensor representation that enables the joint extraction of spatial and temporal features. We further enhance compression with two modules: cross-scale feature aggregation (CFA), which improves spatial representation by fusing coarse and fine features; and bidirectional feature propagation (BiFP), which boosts temporal modelling by enabling mutual feature propagation across frames. Together, these modules exploit spatiotemporal correlations for improved efficiency. Extensive experiments demonstrate that our method consistently outperforms state-of-the-art baselines across various datasets.