DUGAE: Unified Geometry and Attribute Enhancement via Spatiotemporal Correlations for G-PCC Compressed Dynamic Point Clouds

arXiv cs.CV / 3/30/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces DUGAE, a unified post-decoding framework for G-PCC compressed dynamic point clouds that targets the limitations of frame-independent enhancement methods by leveraging inter-frame spatiotemporal correlations.
DUGAE uses three components—DGE-Net for geometry alignment/aggregation via sparse convolution and feature-domain motion compensation, DA-KNN for detail-aware recoloring that improves attribute mapping completeness, and DAE-Net for refined attribute enhancement using temporal feature modeling and attribute motion compensation.
Experiments on seven dynamic point-cloud datasets (from 8iVFB v2, Owlii, and MVUB) show substantial gains over the latest G-PCC geometry-based solid content test model (GeS-TM v10).
For geometry, DUGAE reports an average +11.03 dB BD-PSNR improvement alongside a 93.95% BD-bitrate reduction, and for the luma component it reports +4.23 dB BD-PSNR with a 66.61% BD-bitrate reduction.
The work also improves perceptual quality (PCQM) and reports outperforming V-PCC, with source code planned for release on GitHub.

Abstract

Existing post-decoding quality enhancement methods for point clouds are designed for static data and typically process each frame independently. As a result, they cannot effectively exploit the spatiotemporal correlations present in point cloud sequences.We propose a unified geometry and attribute enhancement framework (DUGAE) for G-PCC compressed dynamic point clouds that explicitly exploits inter-frame spatiotemporal correlations in both geometry and attributes. First, a dynamic geometry enhancement network (DGE-Net) based on sparse convolution (SPConv) and feature-domain geometry motion compensation (GMC) aligns and aggregates spatiotemporal information. Then, a detail-aware k-nearest neighbors (DA-KNN) recoloring module maps the original attributes onto the enhanced geometry at the encoder side, improving mapping completeness and preserving attribute details. Finally, a dynamic attribute enhancement network (DAE-Net) with dedicated temporal feature extraction and feature-domain attribute motion compensation (AMC) refines attributes by modeling complex spatiotemporal correlations. On seven dynamic point clouds from the 8iVFB v2, Owlii, and MVUB datasets, DUGAE significantly enhanced the performance of the latest G-PCC geometry-based solid content test model (GeS-TM v10). For geometry (D1), it achieved an average BD-PSNR gain of 11.03 dB and a 93.95% BD-bitrate reduction. For the luma component, it achieved a 4.23 dB BD-PSNR gain with a 66.61% BD-bitrate reduction. DUGAE also improved perceptual quality (as measured by PCQM) and outperformed V-PCC. Our source code will be released on GitHub at: https://github.com/yuanhui0325/DUGAE