J-SGFT: Joint Spatial and Graph Fourier Domain Learning for Point Cloud Attribute Deblocking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Talha, Muhammad, Yang, Qi, Li, Zhu, Akhtar, Anique, Van Der Auwera, Geert
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911253548498944
author Talha, Muhammad
Yang, Qi
Li, Zhu
Akhtar, Anique
Van Der Auwera, Geert
author_facet Talha, Muhammad
Yang, Qi
Li, Zhu
Akhtar, Anique
Van Der Auwera, Geert
contents Point clouds (PC) are essential for AR/VR and autonomous driving but challenge compression schemes with their size, irregular sampling, and sparsity. MPEG's Geometry-based Point Cloud Compression (GPCC) methods successfully reduce bitrate; however, they introduce significant blocky artifacts in the reconstructed point cloud. We introduce a novel multi-scale postprocessing framework that fuses graph-Fourier latent attribute representations with sparse convolutions and channel-wise attention to efficiently deblock reconstructed point clouds. Against the GPCC TMC13v14 baseline, our approach achieves BD-rate reduction of 18.81\% in the Y channel and 18.14\% in the joint YUV on the 8iVFBv2 dataset, delivering markedly improved visual fidelity with minimal overhead.
format Preprint
id arxiv_https___arxiv_org_abs_2511_05047
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle J-SGFT: Joint Spatial and Graph Fourier Domain Learning for Point Cloud Attribute Deblocking
Talha, Muhammad
Yang, Qi
Li, Zhu
Akhtar, Anique
Van Der Auwera, Geert
Image and Video Processing
Point clouds (PC) are essential for AR/VR and autonomous driving but challenge compression schemes with their size, irregular sampling, and sparsity. MPEG's Geometry-based Point Cloud Compression (GPCC) methods successfully reduce bitrate; however, they introduce significant blocky artifacts in the reconstructed point cloud. We introduce a novel multi-scale postprocessing framework that fuses graph-Fourier latent attribute representations with sparse convolutions and channel-wise attention to efficiently deblock reconstructed point clouds. Against the GPCC TMC13v14 baseline, our approach achieves BD-rate reduction of 18.81\% in the Y channel and 18.14\% in the joint YUV on the 8iVFBv2 dataset, delivering markedly improved visual fidelity with minimal overhead.
title J-SGFT: Joint Spatial and Graph Fourier Domain Learning for Point Cloud Attribute Deblocking
topic Image and Video Processing
url https://arxiv.org/abs/2511.05047