Authentic volumetric avatars from a phone scan
Creating photorealistic avatars of existing people currently requires extensive person-specific data capture, which is usually only accessible to the VFX industry and not the general public. Our work aims to address this drawback by relying only on a short mobile phone capture to obtain a drivable 3...
Gespeichert in:
Veröffentlicht in: | ACM transactions on graphics 2022-07, Vol.41 (4), p.1-19, Article 163 |
---|---|
Hauptverfasser: | , , , , , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 19 |
---|---|
container_issue | 4 |
container_start_page | 1 |
container_title | ACM transactions on graphics |
container_volume | 41 |
creator | Cao, Chen Simon, Tomas Kim, Jin Kyu Schwartz, Gabe Zollhoefer, Michael Saito, Shun-Suke Lombardi, Stephen Wei, Shih-En Belko, Danielle Yu, Shoou-I Sheikh, Yaser Saragih, Jason |
description | Creating photorealistic avatars of existing people currently requires extensive person-specific data capture, which is usually only accessible to the VFX industry and not the general public. Our work aims to address this drawback by relying only on a short mobile phone capture to obtain a drivable 3D head avatar that matches a person's likeness faithfully. In contrast to existing approaches, our architecture avoids the complex task of directly modeling the entire manifold of human appearance, aiming instead to generate an avatar model that can be specialized to novel identities using only small amounts of data. The model dispenses with low-dimensional latent spaces that are commonly employed for hallucinating novel identities, and instead, uses a conditional representation that can extract person-specific information at multiple scales from a high resolution registered neutral phone scan. We achieve high quality results through the use of a novel universal avatar prior that has been trained on high resolution multi-view video captures of facial performances of hundreds of human subjects. By fine-tuning the model using inverse rendering we achieve increased realism and personalize its range of motion. The output of our approach is not only a high-fidelity 3D head avatar that matches the person's facial shape and appearance, but one that can also be driven using a jointly discovered shared global expression space with disentangled controls for gaze direction. Via a series of experiments we demonstrate that our avatars are faithful representations of the subject's likeness. Compared to other state-of-the-art methods for lightweight avatar creation, our approach exhibits superior visual quality and animateability. |
doi_str_mv | 10.1145/3528223.3530143 |
format | Article |
fullrecord | <record><control><sourceid>acm_cross</sourceid><recordid>TN_cdi_crossref_primary_10_1145_3528223_3530143</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3530143</sourcerecordid><originalsourceid>FETCH-LOGICAL-a301t-a6fb386c142d41b2dae3b4bbc36c36edfc173a3ee71cba9396ace505b7b0f1fc3</originalsourceid><addsrcrecordid>eNo9j09Lw0AQxRdRMFbPgqd8gaQzmf3THktRKxS86DnMbnZppUnKblrw2xtpFAbewJv3hp8QjwglolRzUtWiqqgkRYCSrkSGSpnCkF5ciwwMQQGjcyvuUvoCAC2lzkS5Og073w17l5_7w6n1QxxXPvPAMeUh9m3O-XHXdz5Pjrt7cRP4kPzDpDPx-fL8sd4U2_fXt_VqW_D4YihYB0sL7VBWjURbNezJSmsd6XF8ExwaYvLeoLO8pKVm5xUoaywEDI5mYn7pdbFPKfpQH-O-5fhdI9S_uPWEW0-4Y-LpkmDX_h__mT_op1BH</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Authentic volumetric avatars from a phone scan</title><source>ACM Digital Library Complete</source><creator>Cao, Chen ; Simon, Tomas ; Kim, Jin Kyu ; Schwartz, Gabe ; Zollhoefer, Michael ; Saito, Shun-Suke ; Lombardi, Stephen ; Wei, Shih-En ; Belko, Danielle ; Yu, Shoou-I ; Sheikh, Yaser ; Saragih, Jason</creator><creatorcontrib>Cao, Chen ; Simon, Tomas ; Kim, Jin Kyu ; Schwartz, Gabe ; Zollhoefer, Michael ; Saito, Shun-Suke ; Lombardi, Stephen ; Wei, Shih-En ; Belko, Danielle ; Yu, Shoou-I ; Sheikh, Yaser ; Saragih, Jason</creatorcontrib><description>Creating photorealistic avatars of existing people currently requires extensive person-specific data capture, which is usually only accessible to the VFX industry and not the general public. Our work aims to address this drawback by relying only on a short mobile phone capture to obtain a drivable 3D head avatar that matches a person's likeness faithfully. In contrast to existing approaches, our architecture avoids the complex task of directly modeling the entire manifold of human appearance, aiming instead to generate an avatar model that can be specialized to novel identities using only small amounts of data. The model dispenses with low-dimensional latent spaces that are commonly employed for hallucinating novel identities, and instead, uses a conditional representation that can extract person-specific information at multiple scales from a high resolution registered neutral phone scan. We achieve high quality results through the use of a novel universal avatar prior that has been trained on high resolution multi-view video captures of facial performances of hundreds of human subjects. By fine-tuning the model using inverse rendering we achieve increased realism and personalize its range of motion. The output of our approach is not only a high-fidelity 3D head avatar that matches the person's facial shape and appearance, but one that can also be driven using a jointly discovered shared global expression space with disentangled controls for gaze direction. Via a series of experiments we demonstrate that our avatars are faithful representations of the subject's likeness. Compared to other state-of-the-art methods for lightweight avatar creation, our approach exhibits superior visual quality and animateability.</description><identifier>ISSN: 0730-0301</identifier><identifier>EISSN: 1557-7368</identifier><identifier>DOI: 10.1145/3528223.3530143</identifier><language>eng</language><publisher>New York, NY, USA: ACM</publisher><subject>Animation ; Computer graphics ; Computing methodologies</subject><ispartof>ACM transactions on graphics, 2022-07, Vol.41 (4), p.1-19, Article 163</ispartof><rights>Owner/Author</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-a301t-a6fb386c142d41b2dae3b4bbc36c36edfc173a3ee71cba9396ace505b7b0f1fc3</citedby><cites>FETCH-LOGICAL-a301t-a6fb386c142d41b2dae3b4bbc36c36edfc173a3ee71cba9396ace505b7b0f1fc3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://dl.acm.org/doi/pdf/10.1145/3528223.3530143$$EPDF$$P50$$Gacm$$Hfree_for_read</linktopdf><link.rule.ids>314,776,780,2276,27901,27902,40172,75971</link.rule.ids></links><search><creatorcontrib>Cao, Chen</creatorcontrib><creatorcontrib>Simon, Tomas</creatorcontrib><creatorcontrib>Kim, Jin Kyu</creatorcontrib><creatorcontrib>Schwartz, Gabe</creatorcontrib><creatorcontrib>Zollhoefer, Michael</creatorcontrib><creatorcontrib>Saito, Shun-Suke</creatorcontrib><creatorcontrib>Lombardi, Stephen</creatorcontrib><creatorcontrib>Wei, Shih-En</creatorcontrib><creatorcontrib>Belko, Danielle</creatorcontrib><creatorcontrib>Yu, Shoou-I</creatorcontrib><creatorcontrib>Sheikh, Yaser</creatorcontrib><creatorcontrib>Saragih, Jason</creatorcontrib><title>Authentic volumetric avatars from a phone scan</title><title>ACM transactions on graphics</title><addtitle>ACM TOG</addtitle><description>Creating photorealistic avatars of existing people currently requires extensive person-specific data capture, which is usually only accessible to the VFX industry and not the general public. Our work aims to address this drawback by relying only on a short mobile phone capture to obtain a drivable 3D head avatar that matches a person's likeness faithfully. In contrast to existing approaches, our architecture avoids the complex task of directly modeling the entire manifold of human appearance, aiming instead to generate an avatar model that can be specialized to novel identities using only small amounts of data. The model dispenses with low-dimensional latent spaces that are commonly employed for hallucinating novel identities, and instead, uses a conditional representation that can extract person-specific information at multiple scales from a high resolution registered neutral phone scan. We achieve high quality results through the use of a novel universal avatar prior that has been trained on high resolution multi-view video captures of facial performances of hundreds of human subjects. By fine-tuning the model using inverse rendering we achieve increased realism and personalize its range of motion. The output of our approach is not only a high-fidelity 3D head avatar that matches the person's facial shape and appearance, but one that can also be driven using a jointly discovered shared global expression space with disentangled controls for gaze direction. Via a series of experiments we demonstrate that our avatars are faithful representations of the subject's likeness. Compared to other state-of-the-art methods for lightweight avatar creation, our approach exhibits superior visual quality and animateability.</description><subject>Animation</subject><subject>Computer graphics</subject><subject>Computing methodologies</subject><issn>0730-0301</issn><issn>1557-7368</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNo9j09Lw0AQxRdRMFbPgqd8gaQzmf3THktRKxS86DnMbnZppUnKblrw2xtpFAbewJv3hp8QjwglolRzUtWiqqgkRYCSrkSGSpnCkF5ciwwMQQGjcyvuUvoCAC2lzkS5Og073w17l5_7w6n1QxxXPvPAMeUh9m3O-XHXdz5Pjrt7cRP4kPzDpDPx-fL8sd4U2_fXt_VqW_D4YihYB0sL7VBWjURbNezJSmsd6XF8ExwaYvLeoLO8pKVm5xUoaywEDI5mYn7pdbFPKfpQH-O-5fhdI9S_uPWEW0-4Y-LpkmDX_h__mT_op1BH</recordid><startdate>20220722</startdate><enddate>20220722</enddate><creator>Cao, Chen</creator><creator>Simon, Tomas</creator><creator>Kim, Jin Kyu</creator><creator>Schwartz, Gabe</creator><creator>Zollhoefer, Michael</creator><creator>Saito, Shun-Suke</creator><creator>Lombardi, Stephen</creator><creator>Wei, Shih-En</creator><creator>Belko, Danielle</creator><creator>Yu, Shoou-I</creator><creator>Sheikh, Yaser</creator><creator>Saragih, Jason</creator><general>ACM</general><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>20220722</creationdate><title>Authentic volumetric avatars from a phone scan</title><author>Cao, Chen ; Simon, Tomas ; Kim, Jin Kyu ; Schwartz, Gabe ; Zollhoefer, Michael ; Saito, Shun-Suke ; Lombardi, Stephen ; Wei, Shih-En ; Belko, Danielle ; Yu, Shoou-I ; Sheikh, Yaser ; Saragih, Jason</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a301t-a6fb386c142d41b2dae3b4bbc36c36edfc173a3ee71cba9396ace505b7b0f1fc3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Animation</topic><topic>Computer graphics</topic><topic>Computing methodologies</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Cao, Chen</creatorcontrib><creatorcontrib>Simon, Tomas</creatorcontrib><creatorcontrib>Kim, Jin Kyu</creatorcontrib><creatorcontrib>Schwartz, Gabe</creatorcontrib><creatorcontrib>Zollhoefer, Michael</creatorcontrib><creatorcontrib>Saito, Shun-Suke</creatorcontrib><creatorcontrib>Lombardi, Stephen</creatorcontrib><creatorcontrib>Wei, Shih-En</creatorcontrib><creatorcontrib>Belko, Danielle</creatorcontrib><creatorcontrib>Yu, Shoou-I</creatorcontrib><creatorcontrib>Sheikh, Yaser</creatorcontrib><creatorcontrib>Saragih, Jason</creatorcontrib><collection>CrossRef</collection><jtitle>ACM transactions on graphics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Cao, Chen</au><au>Simon, Tomas</au><au>Kim, Jin Kyu</au><au>Schwartz, Gabe</au><au>Zollhoefer, Michael</au><au>Saito, Shun-Suke</au><au>Lombardi, Stephen</au><au>Wei, Shih-En</au><au>Belko, Danielle</au><au>Yu, Shoou-I</au><au>Sheikh, Yaser</au><au>Saragih, Jason</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Authentic volumetric avatars from a phone scan</atitle><jtitle>ACM transactions on graphics</jtitle><stitle>ACM TOG</stitle><date>2022-07-22</date><risdate>2022</risdate><volume>41</volume><issue>4</issue><spage>1</spage><epage>19</epage><pages>1-19</pages><artnum>163</artnum><issn>0730-0301</issn><eissn>1557-7368</eissn><abstract>Creating photorealistic avatars of existing people currently requires extensive person-specific data capture, which is usually only accessible to the VFX industry and not the general public. Our work aims to address this drawback by relying only on a short mobile phone capture to obtain a drivable 3D head avatar that matches a person's likeness faithfully. In contrast to existing approaches, our architecture avoids the complex task of directly modeling the entire manifold of human appearance, aiming instead to generate an avatar model that can be specialized to novel identities using only small amounts of data. The model dispenses with low-dimensional latent spaces that are commonly employed for hallucinating novel identities, and instead, uses a conditional representation that can extract person-specific information at multiple scales from a high resolution registered neutral phone scan. We achieve high quality results through the use of a novel universal avatar prior that has been trained on high resolution multi-view video captures of facial performances of hundreds of human subjects. By fine-tuning the model using inverse rendering we achieve increased realism and personalize its range of motion. The output of our approach is not only a high-fidelity 3D head avatar that matches the person's facial shape and appearance, but one that can also be driven using a jointly discovered shared global expression space with disentangled controls for gaze direction. Via a series of experiments we demonstrate that our avatars are faithful representations of the subject's likeness. Compared to other state-of-the-art methods for lightweight avatar creation, our approach exhibits superior visual quality and animateability.</abstract><cop>New York, NY, USA</cop><pub>ACM</pub><doi>10.1145/3528223.3530143</doi><tpages>19</tpages><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 0730-0301 |
ispartof | ACM transactions on graphics, 2022-07, Vol.41 (4), p.1-19, Article 163 |
issn | 0730-0301 1557-7368 |
language | eng |
recordid | cdi_crossref_primary_10_1145_3528223_3530143 |
source | ACM Digital Library Complete |
subjects | Animation Computer graphics Computing methodologies |
title | Authentic volumetric avatars from a phone scan |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-01T15%3A11%3A30IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-acm_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Authentic%20volumetric%20avatars%20from%20a%20phone%20scan&rft.jtitle=ACM%20transactions%20on%20graphics&rft.au=Cao,%20Chen&rft.date=2022-07-22&rft.volume=41&rft.issue=4&rft.spage=1&rft.epage=19&rft.pages=1-19&rft.artnum=163&rft.issn=0730-0301&rft.eissn=1557-7368&rft_id=info:doi/10.1145/3528223.3530143&rft_dat=%3Cacm_cross%3E3530143%3C/acm_cross%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 |