Book Chapter10.1007/978-981-97-1957-0_10
MPEG AI-Based 3D Graphics Coding Standard
Ge Li,Wei Gao,Guihua Wen +2 more
- 01 Jan 2024
pp 219-241
TL;DR: MPEG AI-Based 3D Graphics Coding Standard explores learning-based point cloud compression techniques presented at the MPEG conference.
read more
Abstract: Previously, our attention was directed toward techniques related to point cloud compression, encompassing transformation, quantization, entropy coding, and others. Within this section, our emphasis shifts toward methods for point cloud compression rooted in deep learning. Moreover, we delve extensively into the realm of learning-based 3D point cloud compression techniques presented at the MPEG conference. This endeavor aims to foster a more profound comprehension of point cloud compression methodologies.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
References
VoxelContext-Net: An Octree based Framework for Point Cloud Compression
Zizheng Que,Guo Lu,Dong Xu +2 more
- 20 Jun 2021
TL;DR: Zhou et al. as discussed by the authors proposed a two-stage deep learning framework called VoxelContext-Net for both static and dynamic point cloud compression, which employs the voxel context to compress the octree structured data.
Multiscale Point Cloud Geometry Compression
Jianqiang Wang,Dandan Ding,Zhu Li,Zhan Ma +3 more
- 23 Mar 2021
TL;DR: In this paper, a multiscale end-to-end learning framework is proposed to hierarchically reconstruct the 3D Point Cloud Geometry (PCG) via progressive re-sampling.
144
OctSqueeze: Octree-Structured Entropy Model for LiDAR Compression
Lila Huang,Shenlong Wang,Kelvin Wong,Jerry Liu,Raquel Urtasun +4 more
- 14 Jun 2020
TL;DR: A novel deep compression algorithm to reduce the memory footprint of LiDAR point clouds and designs a tree-structured conditional entropy model that can be directly applied to octree structures to predict the probability of a symbol’s occurrence.
Surface Representation for Point Clouds
01 Jun 2022
TL;DR: RepSurf as mentioned in this paper is a representative surface representation of point clouds to explicitly depict the very local structure of the point clouds, which can be a plug-and-play module for most point cloud models thanks to its free collaboration with irregular points.
OctAttention: Octree-based Large-scale Contexts Model for Point Cloud Compression
TL;DR: A multiple-contexts deep learning framework called OctAttention employing the octree structure, a memory-efficient representation for point clouds that obtains a 10%-35% BD-Rate gain on the LiDAR benchmark, and saves 95% coding time compared to the voxel-based baseline.