D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning

D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning

Deep learning based 3D reconstruction of single view 2D image is becoming increasingly popular due to their wide range of real-world applications, but this task is inherently challenging because of the partial observability of an object from a single perspective. Recently, state of the art probabili...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ansari, Minhaj Uddin, Bilal, Talha, Akhter, Naeem
Format: Artikel
Sprache:eng
Schlagworte:
Computer Science - Computer Vision and Pattern Recognition
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Ansari, Minhaj Uddin
Bilal, Talha
Akhter, Naeem
description Deep learning based 3D reconstruction of single view 2D image is becoming increasingly popular due to their wide range of real-world applications, but this task is inherently challenging because of the partial observability of an object from a single perspective. Recently, state of the art probability based Occupancy Networks reconstructed 3D surfaces from three different types of input domains: single view 2D image, point cloud and voxel. In this study, we extend the work on Occupancy Networks by exploiting cross-domain learning of image and point cloud domains. Specifically, we first convert the single view 2D image into a simpler point cloud representation, and then reconstruct a 3D surface from it. Our network, the Double Occupancy Network (D-OccNet) outperforms Occupancy Networks in terms of visual quality and details captured in the 3D reconstruction.
doi_str_mv 10.48550/arxiv.2104.13854
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2104_13854</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2104_13854</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-8c2ac4a8411d48b3f8f87cab024534abb3d32ee3f4f14d466b8604376c75962c3</originalsourceid><addsrcrecordid>eNotz81KxDAYheFsXMjoBbgyN5Cany9pdKW0_kFxYJhZly9pIoGZVJIqevfq6OrAuzjwEHIheANWa36F5TN9NFJwaISyGk7Jbc_W3r-E5Yb2YcG0DxNVPd0EP-e6lHe_pDnTXU35lXZlrpX18wFTpkPAkn_qGTmJuK_h_H9XZPtwv-2e2LB-fO7uBoamBWa9RA9oQYgJrFPRRtt6dFyCVoDOqUnJEFSEKGACY5w1HFRrfKuvjfRqRS7_bo-E8a2kA5av8ZcyHinqG0NZQnA</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning</title><source>arXiv.org</source><creator>Ansari, Minhaj Uddin ; Bilal, Talha ; Akhter, Naeem</creator><creatorcontrib>Ansari, Minhaj Uddin ; Bilal, Talha ; Akhter, Naeem</creatorcontrib><description>Deep learning based 3D reconstruction of single view 2D image is becoming increasingly popular due to their wide range of real-world applications, but this task is inherently challenging because of the partial observability of an object from a single perspective. Recently, state of the art probability based Occupancy Networks reconstructed 3D surfaces from three different types of input domains: single view 2D image, point cloud and voxel. In this study, we extend the work on Occupancy Networks by exploiting cross-domain learning of image and point cloud domains. Specifically, we first convert the single view 2D image into a simpler point cloud representation, and then reconstruct a 3D surface from it. Our network, the Double Occupancy Network (D-OccNet) outperforms Occupancy Networks in terms of visual quality and details captured in the 3D reconstruction.</description><identifier>DOI: 10.48550/arxiv.2104.13854</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2021-04</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2104.13854$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2104.13854$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Ansari, Minhaj Uddin</creatorcontrib><creatorcontrib>Bilal, Talha</creatorcontrib><creatorcontrib>Akhter, Naeem</creatorcontrib><title>D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning</title><description>Deep learning based 3D reconstruction of single view 2D image is becoming increasingly popular due to their wide range of real-world applications, but this task is inherently challenging because of the partial observability of an object from a single perspective. Recently, state of the art probability based Occupancy Networks reconstructed 3D surfaces from three different types of input domains: single view 2D image, point cloud and voxel. In this study, we extend the work on Occupancy Networks by exploiting cross-domain learning of image and point cloud domains. Specifically, we first convert the single view 2D image into a simpler point cloud representation, and then reconstruct a 3D surface from it. Our network, the Double Occupancy Network (D-OccNet) outperforms Occupancy Networks in terms of visual quality and details captured in the 3D reconstruction.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz81KxDAYheFsXMjoBbgyN5Cany9pdKW0_kFxYJhZly9pIoGZVJIqevfq6OrAuzjwEHIheANWa36F5TN9NFJwaISyGk7Jbc_W3r-E5Yb2YcG0DxNVPd0EP-e6lHe_pDnTXU35lXZlrpX18wFTpkPAkn_qGTmJuK_h_H9XZPtwv-2e2LB-fO7uBoamBWa9RA9oQYgJrFPRRtt6dFyCVoDOqUnJEFSEKGACY5w1HFRrfKuvjfRqRS7_bo-E8a2kA5av8ZcyHinqG0NZQnA</recordid><startdate>20210428</startdate><enddate>20210428</enddate><creator>Ansari, Minhaj Uddin</creator><creator>Bilal, Talha</creator><creator>Akhter, Naeem</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20210428</creationdate><title>D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning</title><author>Ansari, Minhaj Uddin ; Bilal, Talha ; Akhter, Naeem</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-8c2ac4a8411d48b3f8f87cab024534abb3d32ee3f4f14d466b8604376c75962c3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Ansari, Minhaj Uddin</creatorcontrib><creatorcontrib>Bilal, Talha</creatorcontrib><creatorcontrib>Akhter, Naeem</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Ansari, Minhaj Uddin</au><au>Bilal, Talha</au><au>Akhter, Naeem</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning</atitle><date>2021-04-28</date><risdate>2021</risdate><abstract>Deep learning based 3D reconstruction of single view 2D image is becoming increasingly popular due to their wide range of real-world applications, but this task is inherently challenging because of the partial observability of an object from a single perspective. Recently, state of the art probability based Occupancy Networks reconstructed 3D surfaces from three different types of input domains: single view 2D image, point cloud and voxel. In this study, we extend the work on Occupancy Networks by exploiting cross-domain learning of image and point cloud domains. Specifically, we first convert the single view 2D image into a simpler point cloud representation, and then reconstruct a 3D surface from it. Our network, the Double Occupancy Network (D-OccNet) outperforms Occupancy Networks in terms of visual quality and details captured in the 3D reconstruction.</abstract><doi>10.48550/arxiv.2104.13854</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2104.13854
ispartof
issn
language eng
recordid cdi_arxiv_primary_2104_13854
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
title D-OccNet: Detailed 3D Reconstruction Using Cross-Domain Learning
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-27T04%3A45%3A00IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=D-OccNet:%20Detailed%203D%20Reconstruction%20Using%20Cross-Domain%20Learning&rft.au=Ansari,%20Minhaj%20Uddin&rft.date=2021-04-28&rft_id=info:doi/10.48550/arxiv.2104.13854&rft_dat=%3Carxiv_GOX%3E2104_13854%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true