Loading…

Deep Learning for Robust Normal Estimation in Unstructured Point Clouds

Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density i...

Full description

Saved in:
Bibliographic Details
Published in:Computer graphics forum 2016-08, Vol.35 (5), p.281-290
Main Authors: Boulch, Alexandre, Marlet, Renaud
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3
cites cdi_FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3
container_end_page 290
container_issue 5
container_start_page 281
container_title Computer graphics forum
container_volume 35
creator Boulch, Alexandre
Marlet, Renaud
description Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density is not uniform, to prevent bias. Rather than resorting to manually‐designed geometric priors, we propose to learn how to make these decisions, using ground‐truth data made from synthetic scenes. For this, we project a discretized Hough space representing normal directions onto a structure amenable to deep learning. The resulting normal estimation method outperforms most of the time the state of the art regarding robustness to outliers, to noise and to point density variation, in the presence of sharp edges, while remaining fast, scaling up to millions of points.
doi_str_mv 10.1111/cgf.12983
format article
fullrecord <record><control><sourceid>proquest_hal_p</sourceid><recordid>TN_cdi_hal_primary_oai_HAL_hal_01631736v1</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>4147803231</sourcerecordid><originalsourceid>FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3</originalsourceid><addsrcrecordid>eNp1kMtKAzEUhoMoWC8L3yDgRhfTJpPbzFKqVmGoRVoENyHNZDQ6ndRkRu3bm1ovIJhNwuH7c875ADjCqI_jGeiHqo_TPCNboIcpF0nGWb4NegjHt0CM7YK9EJ4QQlRw1gOjc2OWsDDKN7Z5gJXz8NbNu9DCsfMLVcOL0NqFaq1roG3grAmt73TbeVPCibNNC4e168pwAHYqVQdz-HXvg9nlxXR4lRQ3o-vhWZFohgRJOBcE0wyrMkda5RmrSpMKUmap4JqWvKR4zhkTLKMqJRoZTllKKzTPEDdsrsk-ON38-6hqufRxNL-STll5dVbIdS0uSrAg_BVH9mTDLr176Uxo5cIGbepaNcZ1QeKMMMYxZXlEj_-gT67zTdwkUjiKpBiT3-bauxC8qX4mwEiu9cuoX37qj-xgw77Z2qz-B-VwdPmdSDYJG1rz_pNQ_llGa4LJu_FIFtMJIuP7Qk7IB6U7kns</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>1811464113</pqid></control><display><type>article</type><title>Deep Learning for Robust Normal Estimation in Unstructured Point Clouds</title><source>Business Source Ultimate【Trial: -2024/12/31】【Remote access available】</source><source>Wiley</source><source>Art &amp; Architecture Source</source><creator>Boulch, Alexandre ; Marlet, Renaud</creator><creatorcontrib>Boulch, Alexandre ; Marlet, Renaud</creatorcontrib><description>Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density is not uniform, to prevent bias. Rather than resorting to manually‐designed geometric priors, we propose to learn how to make these decisions, using ground‐truth data made from synthetic scenes. For this, we project a discretized Hough space representing normal directions onto a structure amenable to deep learning. The resulting normal estimation method outperforms most of the time the state of the art regarding robustness to outliers, to noise and to point density variation, in the presence of sharp edges, while remaining fast, scaling up to millions of points.</description><identifier>ISSN: 0167-7055</identifier><identifier>EISSN: 1467-8659</identifier><identifier>DOI: 10.1111/cgf.12983</identifier><language>eng</language><publisher>Oxford: Blackwell Publishing Ltd</publisher><subject>Algorithms ; Analysis ; Computer graphics ; Computer Science ; Decisions ; Deep learning ; Density ; Estimating techniques ; Image processing systems ; Learning ; Rendering ; Robustness ; Sampling ; Scaling up ; Studies ; Three dimensional models</subject><ispartof>Computer graphics forum, 2016-08, Vol.35 (5), p.281-290</ispartof><rights>2016 The Author(s) Computer Graphics Forum © 2016 The Eurographics Association and John Wiley &amp; Sons Ltd. Published by John Wiley &amp; Sons Ltd.</rights><rights>2016 The Eurographics Association and John Wiley &amp; Sons Ltd.</rights><rights>Distributed under a Creative Commons Attribution 4.0 International License</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3</citedby><cites>FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3</cites><orcidid>0000-0002-4196-9665 ; 0000-0003-1612-1758</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>230,314,780,784,885,27924,27925</link.rule.ids><backlink>$$Uhttps://hal.science/hal-01631736$$DView record in HAL$$Hfree_for_read</backlink></links><search><creatorcontrib>Boulch, Alexandre</creatorcontrib><creatorcontrib>Marlet, Renaud</creatorcontrib><title>Deep Learning for Robust Normal Estimation in Unstructured Point Clouds</title><title>Computer graphics forum</title><addtitle>Computer Graphics Forum</addtitle><description>Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density is not uniform, to prevent bias. Rather than resorting to manually‐designed geometric priors, we propose to learn how to make these decisions, using ground‐truth data made from synthetic scenes. For this, we project a discretized Hough space representing normal directions onto a structure amenable to deep learning. The resulting normal estimation method outperforms most of the time the state of the art regarding robustness to outliers, to noise and to point density variation, in the presence of sharp edges, while remaining fast, scaling up to millions of points.</description><subject>Algorithms</subject><subject>Analysis</subject><subject>Computer graphics</subject><subject>Computer Science</subject><subject>Decisions</subject><subject>Deep learning</subject><subject>Density</subject><subject>Estimating techniques</subject><subject>Image processing systems</subject><subject>Learning</subject><subject>Rendering</subject><subject>Robustness</subject><subject>Sampling</subject><subject>Scaling up</subject><subject>Studies</subject><subject>Three dimensional models</subject><issn>0167-7055</issn><issn>1467-8659</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2016</creationdate><recordtype>article</recordtype><recordid>eNp1kMtKAzEUhoMoWC8L3yDgRhfTJpPbzFKqVmGoRVoENyHNZDQ6ndRkRu3bm1ovIJhNwuH7c875ADjCqI_jGeiHqo_TPCNboIcpF0nGWb4NegjHt0CM7YK9EJ4QQlRw1gOjc2OWsDDKN7Z5gJXz8NbNu9DCsfMLVcOL0NqFaq1roG3grAmt73TbeVPCibNNC4e168pwAHYqVQdz-HXvg9nlxXR4lRQ3o-vhWZFohgRJOBcE0wyrMkda5RmrSpMKUmap4JqWvKR4zhkTLKMqJRoZTllKKzTPEDdsrsk-ON38-6hqufRxNL-STll5dVbIdS0uSrAg_BVH9mTDLr176Uxo5cIGbepaNcZ1QeKMMMYxZXlEj_-gT67zTdwkUjiKpBiT3-bauxC8qX4mwEiu9cuoX37qj-xgw77Z2qz-B-VwdPmdSDYJG1rz_pNQ_llGa4LJu_FIFtMJIuP7Qk7IB6U7kns</recordid><startdate>201608</startdate><enddate>201608</enddate><creator>Boulch, Alexandre</creator><creator>Marlet, Renaud</creator><general>Blackwell Publishing Ltd</general><general>Wiley</general><scope>BSCLL</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>F28</scope><scope>FR3</scope><scope>1XC</scope><scope>VOOES</scope><orcidid>https://orcid.org/0000-0002-4196-9665</orcidid><orcidid>https://orcid.org/0000-0003-1612-1758</orcidid></search><sort><creationdate>201608</creationdate><title>Deep Learning for Robust Normal Estimation in Unstructured Point Clouds</title><author>Boulch, Alexandre ; Marlet, Renaud</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2016</creationdate><topic>Algorithms</topic><topic>Analysis</topic><topic>Computer graphics</topic><topic>Computer Science</topic><topic>Decisions</topic><topic>Deep learning</topic><topic>Density</topic><topic>Estimating techniques</topic><topic>Image processing systems</topic><topic>Learning</topic><topic>Rendering</topic><topic>Robustness</topic><topic>Sampling</topic><topic>Scaling up</topic><topic>Studies</topic><topic>Three dimensional models</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Boulch, Alexandre</creatorcontrib><creatorcontrib>Marlet, Renaud</creatorcontrib><collection>Istex</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>ANTE: Abstracts in New Technology &amp; Engineering</collection><collection>Engineering Research Database</collection><collection>Hyper Article en Ligne (HAL)</collection><collection>Hyper Article en Ligne (HAL) (Open Access)</collection><jtitle>Computer graphics forum</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Boulch, Alexandre</au><au>Marlet, Renaud</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Deep Learning for Robust Normal Estimation in Unstructured Point Clouds</atitle><jtitle>Computer graphics forum</jtitle><addtitle>Computer Graphics Forum</addtitle><date>2016-08</date><risdate>2016</risdate><volume>35</volume><issue>5</issue><spage>281</spage><epage>290</epage><pages>281-290</pages><issn>0167-7055</issn><eissn>1467-8659</eissn><abstract>Normal estimation in point clouds is a crucial first step for numerous algorithms, from surface reconstruction and scene understanding to rendering. A recurrent issue when estimating normals is to make appropriate decisions close to sharp features, not to smooth edges, or when the sampling density is not uniform, to prevent bias. Rather than resorting to manually‐designed geometric priors, we propose to learn how to make these decisions, using ground‐truth data made from synthetic scenes. For this, we project a discretized Hough space representing normal directions onto a structure amenable to deep learning. The resulting normal estimation method outperforms most of the time the state of the art regarding robustness to outliers, to noise and to point density variation, in the presence of sharp edges, while remaining fast, scaling up to millions of points.</abstract><cop>Oxford</cop><pub>Blackwell Publishing Ltd</pub><doi>10.1111/cgf.12983</doi><tpages>10</tpages><orcidid>https://orcid.org/0000-0002-4196-9665</orcidid><orcidid>https://orcid.org/0000-0003-1612-1758</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 0167-7055
ispartof Computer graphics forum, 2016-08, Vol.35 (5), p.281-290
issn 0167-7055
1467-8659
language eng
recordid cdi_hal_primary_oai_HAL_hal_01631736v1
source Business Source Ultimate【Trial: -2024/12/31】【Remote access available】; Wiley; Art & Architecture Source
subjects Algorithms
Analysis
Computer graphics
Computer Science
Decisions
Deep learning
Density
Estimating techniques
Image processing systems
Learning
Rendering
Robustness
Sampling
Scaling up
Studies
Three dimensional models
title Deep Learning for Robust Normal Estimation in Unstructured Point Clouds
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T14%3A22%3A47IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_hal_p&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Deep%20Learning%20for%20Robust%20Normal%20Estimation%20in%20Unstructured%20Point%20Clouds&rft.jtitle=Computer%20graphics%20forum&rft.au=Boulch,%20Alexandre&rft.date=2016-08&rft.volume=35&rft.issue=5&rft.spage=281&rft.epage=290&rft.pages=281-290&rft.issn=0167-7055&rft.eissn=1467-8659&rft_id=info:doi/10.1111/cgf.12983&rft_dat=%3Cproquest_hal_p%3E4147803231%3C/proquest_hal_p%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c5073-66731481ad90ca985fde273d8276c4d6d41b6557584a23c0e64524f0b806e5bc3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=1811464113&rft_id=info:pmid/&rfr_iscdi=true