Lossless Point Cloud Attribute Compression with Normal-based Intra Prediction

TL;DR

This paper introduces a normal-based intra prediction method for lossless point cloud attribute compression, leveraging normals to improve compression efficiency over traditional distance-based methods in G-PCC.

Contribution

The paper proposes a novel normal-based intra prediction scheme that enhances lossless attribute compression by utilizing point cloud normals for better predictor selection.

Findings

Achieves 2.1% average gain in lossless attribute coding

Outperforms G-PCC anchor in experimental tests

Utilizes normal angles for improved local similarity exploration

Abstract

The sparse LiDAR point clouds become more and more popular in various applications, e.g., the autonomous driving. However, for this type of data, there exists much under-explored space in the corresponding compression framework proposed by MPEG, i.e., geometry-based point cloud compression (G-PCC). In G-PCC, only the distance-based similarity is considered in the intra prediction for the attribute compression. In this paper, we propose a normal-based intra prediction scheme, which provides a more efficient lossless attribute compression by introducing the normals of point clouds. The angle between normals is used to further explore accurate local similarity, which optimizes the selection of predictors. We implement our method into the G-PCC reference software. Experimental results over LiDAR acquired datasets demonstrate that our proposed method is able to deliver better compression performance than the G-PCC anchor, with $2.1\%$ gains on average for lossless attribute coding.