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:20250110T023313Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T131400
DTEND;TZID=Asia/Tokyo:20241206T132800
UID:siggraphasia_SIGGRAPH Asia 2024_sess145_papers_544@linklings.com
SUMMARY:Gaussian Opacity Fields: Efficient Adaptive Surface Reconstruction
  in Unbounded Scenes
DESCRIPTION:Technical Papers\n\nZehao Yu (University of Tübingen, Tübingen
  AI Center); Torsten Sattler (Czech Technical University in Prague); and A
 ndreas Geiger (University of Tübingen, Tübingen AI Center)\n\nRecently, 3D
  Gaussian Splatting (3DGS) has demonstrated impressive novel view synthesi
 s results, while allowing the rendering of high-resolution images in real-
 time. However, leveraging 3D Gaussians for surface reconstruction poses si
 gnificant challenges due to the explicit and disconnected nature of 3D Gau
 ssians. In this work, we present Gaussian Opacity Fields (GOF), a novel ap
 proach for efficient, high-quality, and adaptive surface reconstruction in
  unbounded scenes. Our GOF is derived from ray-tracing-based volume render
 ing of 3D Gaussians, enabling direct geometry extraction from 3D Gaussians
  by identifying its levelset, without resorting to Poisson reconstruction 
 or TSDF fusion as in previous work. We approximate the surface normal of G
 aussians as the normal of the ray-Gaussian intersection plane, enabling th
 e application of regularization that significantly enhances geometry. Furt
 hermore, we develop an efficient geometry extraction method utilizing Marc
 hing Tetrahedra, where the tetrahedral grids are induced from 3D Gaussians
  and thus adapt to the scene's complexity. Our evaluations reveal that GOF
  surpasses existing 3DGS-based methods in surface reconstruction and novel
  view synthesis. Further, it compares favorably to or even outperforms, ne
 ural implicit methods in both quality and speed.\n\nRegistration Category:
  Full Access, Full Access Supporter\n\nLanguage Format: English Language\n
 \nSession Chair: Hao (Richard) Zhang (Simon Fraser University, Amazon)
URL:https://asia.siggraph.org/2024/program/?id=papers_544&sess=sess145
END:VEVENT
END:VCALENDAR