八叉树结构下动态点云的无损几何编码方法

Zhecheng Wang; Shuai Wan; Lei Wei; Fuzheng Yang

doi:10.7652/xjtuxb202312002

八叉树结构下动态点云的无损几何编码方法

Translated title of the contribution: Lossless Geometry Coding Method for Dynamic Point Clouds under Octree Structure

Zhecheng Wang
, Shuai Wan
, Lei Wei
, Fuzheng Yang

School of Electronics and Information

Northwestern Polytechnical University Xian
Royal Melbourne Institute of Technology University
Xidian University

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

Tofurtherimprovetheperformanceoflosslesscompressionofdynamicpointclouds a methodofgeometryinformationentropycodingunderanoctreestructurewasproposed.Bythis method forintra-framespatialcorrelation thecontextsofintra-frameneighbornodesandthe contextsofintra-frameneighborparentnodesareestablishedbasedonthecodedneighborhood informationofthecurrentoctreenode^;forinter-frametemporalcorrelation thepreviouslycoded pointcloudframeisemployedasthereferenceframe andtheoctreenodeinsidethereference frameatthesamepositionasthecurrentoctreenodeisusedasthereferencenode.Theinterframecontextismodeledusingthereferencenodeanditsparentnode.Tofullyutilizethemodeledcontextsandaccuratelyestimatetheconditionalprobabilitythatthecurrentnodeisnon-emptyunderdifferentcontexts asecond-levelprobabilityestimation methodbasedontheexponential movingaveragewasproposed.Bythismethod theprobabilityisfirstestimatedunderthecontextofintra-frameneighbornodesandthecontextofintra-frameneighborparentnodes respectively.Then asetofsecond-levelcontextsaremodeledusingtheinter-framecontextsandthe resultsoftheprobabilityestimation andtheprobabilityisestimatedagainunderthesecond-level contexts.Intheend thelosslesscompressionisrealizedbythebinaryarithmeticencoder.To evaluatethecompressionperformanceoftheproposed method thecommonlyused Microsoft voxelizedupperbodies(MVUB)and8ivoxelizedfullbodies(8iVFB)datasetswereselectedfor performancetesting.Theexperimentalresultsshowthattheproposedmethodhasahigherlosslesscompressionperformancethanthemethodsdevelopedrecently Withaveragecodinggainsof 2.2%to28.7%.

Translated title of the contribution	Lossless Geometry Coding Method for Dynamic Point Clouds under Octree Structure
Original language	Chinese (Traditional)
Pages (from-to)	11-19
Number of pages	9
Journal	Hsi-An Chiao Tung Ta Hsueh/Journal of Xi'an Jiaotong University
Volume	57
Issue number	12
DOIs	https://doi.org/10.7652/xjtuxb202312002
State	Published - Dec 2023

Access to Document

10.7652/xjtuxb202312002

Cite this

@article{c466fb6360174e50a8e8ede4a9840003,

title = "八叉树结构下动态点云的无损几何编码方法",

abstract = "Tofurtherimprovetheperformanceoflosslesscompressionofdynamicpointclouds a methodofgeometryinformationentropycodingunderanoctreestructurewasproposed.Bythis method forintra-framespatialcorrelation thecontextsofintra-frameneighbornodesandthe contextsofintra-frameneighborparentnodesareestablishedbasedonthecodedneighborhood informationofthecurrentoctreenode;forinter-frametemporalcorrelation thepreviouslycoded pointcloudframeisemployedasthereferenceframe andtheoctreenodeinsidethereference frameatthesamepositionasthecurrentoctreenodeisusedasthereferencenode.Theinterframecontextismodeledusingthereferencenodeanditsparentnode.Tofullyutilizethemodeledcontextsandaccuratelyestimatetheconditionalprobabilitythatthecurrentnodeisnon-emptyunderdifferentcontexts asecond-levelprobabilityestimation methodbasedontheexponential movingaveragewasproposed.Bythismethod theprobabilityisfirstestimatedunderthecontextofintra-frameneighbornodesandthecontextofintra-frameneighborparentnodes respectively.Then asetofsecond-levelcontextsaremodeledusingtheinter-framecontextsandthe resultsoftheprobabilityestimation andtheprobabilityisestimatedagainunderthesecond-level contexts.Intheend thelosslesscompressionisrealizedbythebinaryarithmeticencoder.To evaluatethecompressionperformanceoftheproposed method thecommonlyused Microsoft voxelizedupperbodies(MVUB)and8ivoxelizedfullbodies(8iVFB)datasetswereselectedfor performancetesting.Theexperimentalresultsshowthattheproposedmethodhasahigherlosslesscompressionperformancethanthemethodsdevelopedrecently Withaveragecodinggainsof 2.2\%to28.7\%.",

keywords = "arithmetic coding, entropy coding, octree, point cloud compression, probability estimation",

author = "Zhecheng Wang and Shuai Wan and Lei Wei and Fuzheng Yang",

year = "2023",

month = dec,

doi = "10.7652/xjtuxb202312002",

language = "繁体中文",

volume = "57",

pages = "11--19",

journal = "Hsi-An Chiao Tung Ta Hsueh/Journal of Xi'an Jiaotong University",

issn = "0253-987X",

publisher = "Xi'an Jiaotong University",

number = "12",

}

TY - JOUR

T1 - 八叉树结构下动态点云的无损几何编码方法

AU - Wang, Zhecheng

AU - Wan, Shuai

AU - Wei, Lei

AU - Yang, Fuzheng

PY - 2023/12

Y1 - 2023/12

N2 - Tofurtherimprovetheperformanceoflosslesscompressionofdynamicpointclouds a methodofgeometryinformationentropycodingunderanoctreestructurewasproposed.Bythis method forintra-framespatialcorrelation thecontextsofintra-frameneighbornodesandthe contextsofintra-frameneighborparentnodesareestablishedbasedonthecodedneighborhood informationofthecurrentoctreenode;forinter-frametemporalcorrelation thepreviouslycoded pointcloudframeisemployedasthereferenceframe andtheoctreenodeinsidethereference frameatthesamepositionasthecurrentoctreenodeisusedasthereferencenode.Theinterframecontextismodeledusingthereferencenodeanditsparentnode.Tofullyutilizethemodeledcontextsandaccuratelyestimatetheconditionalprobabilitythatthecurrentnodeisnon-emptyunderdifferentcontexts asecond-levelprobabilityestimation methodbasedontheexponential movingaveragewasproposed.Bythismethod theprobabilityisfirstestimatedunderthecontextofintra-frameneighbornodesandthecontextofintra-frameneighborparentnodes respectively.Then asetofsecond-levelcontextsaremodeledusingtheinter-framecontextsandthe resultsoftheprobabilityestimation andtheprobabilityisestimatedagainunderthesecond-level contexts.Intheend thelosslesscompressionisrealizedbythebinaryarithmeticencoder.To evaluatethecompressionperformanceoftheproposed method thecommonlyused Microsoft voxelizedupperbodies(MVUB)and8ivoxelizedfullbodies(8iVFB)datasetswereselectedfor performancetesting.Theexperimentalresultsshowthattheproposedmethodhasahigherlosslesscompressionperformancethanthemethodsdevelopedrecently Withaveragecodinggainsof 2.2%to28.7%.

AB - Tofurtherimprovetheperformanceoflosslesscompressionofdynamicpointclouds a methodofgeometryinformationentropycodingunderanoctreestructurewasproposed.Bythis method forintra-framespatialcorrelation thecontextsofintra-frameneighbornodesandthe contextsofintra-frameneighborparentnodesareestablishedbasedonthecodedneighborhood informationofthecurrentoctreenode;forinter-frametemporalcorrelation thepreviouslycoded pointcloudframeisemployedasthereferenceframe andtheoctreenodeinsidethereference frameatthesamepositionasthecurrentoctreenodeisusedasthereferencenode.Theinterframecontextismodeledusingthereferencenodeanditsparentnode.Tofullyutilizethemodeledcontextsandaccuratelyestimatetheconditionalprobabilitythatthecurrentnodeisnon-emptyunderdifferentcontexts asecond-levelprobabilityestimation methodbasedontheexponential movingaveragewasproposed.Bythismethod theprobabilityisfirstestimatedunderthecontextofintra-frameneighbornodesandthecontextofintra-frameneighborparentnodes respectively.Then asetofsecond-levelcontextsaremodeledusingtheinter-framecontextsandthe resultsoftheprobabilityestimation andtheprobabilityisestimatedagainunderthesecond-level contexts.Intheend thelosslesscompressionisrealizedbythebinaryarithmeticencoder.To evaluatethecompressionperformanceoftheproposed method thecommonlyused Microsoft voxelizedupperbodies(MVUB)and8ivoxelizedfullbodies(8iVFB)datasetswereselectedfor performancetesting.Theexperimentalresultsshowthattheproposedmethodhasahigherlosslesscompressionperformancethanthemethodsdevelopedrecently Withaveragecodinggainsof 2.2%to28.7%.

KW - arithmetic coding

KW - entropy coding

KW - octree

KW - point cloud compression

KW - probability estimation

UR - https://www.scopus.com/pages/publications/85180972531

U2 - 10.7652/xjtuxb202312002

DO - 10.7652/xjtuxb202312002

M3 - 文章

AN - SCOPUS:85180972531

SN - 0253-987X

VL - 57

SP - 11

EP - 19

JO - Hsi-An Chiao Tung Ta Hsueh/Journal of Xi'an Jiaotong University

JF - Hsi-An Chiao Tung Ta Hsueh/Journal of Xi'an Jiaotong University

IS - 12

ER -

八叉树结构下动态点云的无损几何编码方法

Abstract

Access to Document

Other files and links

Fingerprint

Cite this