BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035306Z
LOCATION:Auditorium\, Level 5\, West Wing
DTSTART;TZID=Asia/Seoul:20221206T100000
DTEND;TZID=Asia/Seoul:20221206T120000
UID:siggraphasia_SIGGRAPH Asia 2022_sess153_papers_228@linklings.com
SUMMARY:DeepMVSHair: Deep Hair Modeling from Sparse Views
DESCRIPTION:Technical Papers\n\nDeepMVSHair: Deep Hair Modeling from Spars
 e Views\n\nKuang, Chen, Fu, Zhou, Zheng\n\nWe present DeepMVSHair, the fir
 st deep learning-based method for multi-view hair strand reconstruction. T
 he key component of our pipeline is HairMVSNet, a differentiable neural ar
 chitecture which represents a spatial hair structure as a continuous 3D ha
 ir growing direction field implicitly. Specifically, given a 3D query poin
 t, we decide its occupancy (whether it is inside the hair volume) and dire
 ction from observed 2D structure features. With the query point’s pixel-al
 igned features from each input view, we utilize a view-aware transformer e
 ncoder to aggregate anisotropic structure features to an integrated repres
 entation, which is decoded to yield\n3D occupancy and direction at the que
 ry point. HairMVSNet effectively gathers multi-view hair structure feature
 s and preserves high-frequency details based on this implicit representati
 on. Guided by HairMVSNet, our hair-growing algorithm produces results fait
 hful to input multi-view images. We propose a novel image-guided multi-vie
 w strand deformation algorithm to enrich modeling details further. Extensi
 ve experiments show that the results by our sparse-view method are compara
 ble to those by state-of-the-art dense multi-view methods and significantl
 y better than those by single-view and sparse-view methods. In addition, o
 ur method is an\norder of magnitude faster than previous multi-view hair m
 odeling methods.\n\nRegistration Category: FULL ACCESS, EXPERIENCE PLUS AC
 CESS, EXPERIENCE ACCESS, TRADE EXHIBITOR\n\nLanguage: ENGLISH\n\nFormat: I
 N-PERSON
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_228&sess=sess15
 3
END:VEVENT
END:VCALENDAR