Learning-based Lossless Point Cloud Geometry Coding using Sparse Tensors

TL;DR

This paper introduces a novel lossless point cloud geometry compression method that directly processes point data using sparse neural networks, achieving significant rate savings over existing standards.

Contribution

It presents a new context-based compression approach operating directly on point clouds with sparse convolutions, reducing computational cost and improving compression efficiency.

Findings

Outperforms MPEG standard with 52% average rate savings

Operates directly on point data, preserving geometric correlations

Reduces computational cost compared to voxel/octree methods

Abstract

Most point cloud compression methods operate in the voxel or octree domain which is not the original representation of point clouds. Those representations either remove the geometric information or require high computational power for processing. In this paper, we propose a context-based lossless point cloud geometry compression that directly processes the point representation. Operating on a point representation allows us to preserve geometry correlation between points and thus to obtain an accurate context model while significantly reduce the computational cost. Specifically, our method uses a sparse convolution neural network to estimate the voxel occupancy sequentially from the x,y,z input data. Experimental results show that our method outperforms the state-of-the-art geometry compression standard from MPEG with average rate savings of 52% on a diverse set of point clouds from four different datasets.