BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250110T023312Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241204T131100
DTEND;TZID=Asia/Tokyo:20241204T132300
UID:siggraphasia_SIGGRAPH Asia 2024_sess115_papers_678@linklings.com
SUMMARY:3DGSR: Implicit Surface Reconstruction with 3D Gaussian Splatting
DESCRIPTION:Technical Papers\n\nXiaoyang Lyu, Yang-Tian Sun, Yi-Hua Huang,
  Xiuzhe Wu, and Ziyi Yang (University of Hong Kong); Yilun Chen and Jiangm
 iao Pang (Shanghai Artificial Intelligence Laboratory); and Xiaojuan Qi (U
 niversity of Hong Kong)\n\nIn this paper, we present an implicit surface r
 econstruction method with\n3D Gaussian Splatting (3DGS), namely 3DGSR, tha
 t allows for accurate 3D\nreconstruction with intricate details while inhe
 riting the high efficiency and\nrendering quality of 3DGS. The key insight
  is to incorporate an implicit\nsigned distance field (SDF) within 3D Gaus
 sians for surface modeling and to\nenable the alignment and joint optimiza
 tion of both SDF and 3D Gaussians.\nTo achieve this, we design coupling st
 rategies that align and associate the\nSDF with 3D Gaussians, allowing for
  unified optimization and enforcing\nsurface constraints on the 3D Gaussia
 ns. With alignment, optimizing the 3D\nGaussians provides supervisory sign
 als for SDF learning, enabling the recon-\nstruction of intricate details.
  However, this only offers sparse supervisory\nsignals to the SDF at locat
 ions occupied by Gaussians, which is insufficient\nfor learning a continuo
 us SDF. Then, to address this limitation, we incor-\nporate volumetric ren
 dering and align the rendered geometric attributes\n(depth, normal) with t
 hat derived from 3DGS. In sum, these two designs\nallow SDF and 3DGS to be
  aligned, jointly optimized, and mutually boosted.\nOur extensive experime
 ntal results demonstrate that our 3DGSR enables\nhigh-quality 3D surface r
 econstruction while preserving the efficiency and\nrendering quality of 3D
 GS. Besides, our method competes favorably with\nleading surface reconstru
 ction techniques while offering a more efficient\nlearning process and muc
 h better rendering qualities.\n\nRegistration Category: Full Access, Full 
 Access Supporter\n\nLanguage Format: English Language\n\nSession Chair: Pe
 ng-Shuai Wang (Peking University)
URL:https://asia.siggraph.org/2024/program/?id=papers_678&sess=sess115
END:VEVENT
END:VCALENDAR