BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030653Z
LOCATION:Meeting Room S421\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251216T170200
DTEND;TZID=Asia/Hong_Kong:20251216T171300
UID:siggraphasia_SIGGRAPH Asia 2025_sess128_papers_1830@linklings.com
SUMMARY:SOF: Sorted Opacity Fields for Fast Unbounded Surface Reconstructi
 on
DESCRIPTION:Lukas Radl and Felix Windisch (Graz University of Technology);
  Thomas Deixelberger and Jozef Hladky (Huawei Technologies); Michael Stein
 er (Graz University of Technology); Dieter Schmalstieg (Graz University of
  Technology, University of Stuttgart); and Markus Steinberger (Graz Univer
 sity of Technology, Huawei Technologies)\n\nRecent advances in 3D Gaussian
  representations have significantly improved the quality and efficiency of
  image-based scene reconstruction. \nTheir explicit nature facilitates rea
 l-time rendering and fast optimization, yet extracting accurate surfaces -
  particularly in large-scale, unbounded environments - remains a difficult
  task. \nMany existing methods rely on approximate depth estimates and glo
 bal sorting heuristics, which can introduce artifacts and limit the fideli
 ty of the reconstructed mesh.\nIn this paper, we present Sorted Opacity Fi
 elds (SOF), a method designed to recover detailed surfaces from 3D Gaussia
 ns with both speed and precision. \nOur approach improves upon prior work 
 by introducing hierarchical resorting and a robust formulation of Gaussian
  depth, which better aligns with the level-set. \nTo enhance mesh quality,
  we incorporate a level-set regularizer operating on the opacity field and
  introduce losses that encourage geometrically-consistent primitive shapes
 . \nIn addition, we develop a parallelized Marching Tetrahedra algorithm t
 ailored to our opacity formulation, reducing meshing time by up to an orde
 r of magnitude.\nAs demonstrated by our quantitative evaluation, SOF achie
 ves higher reconstruction accuracy while cutting total processing time by 
 more than a factor of three.\nThese results mark a step forward in turning
  efficient Gaussian-based rendering into equally efficient geometry extrac
 tion.\n\nRegistration Category: Full Access, Full Access Supporter\n\nSess
 ion Chair: Peng-Shuai Wang (Peking University)\n\n
END:VEVENT
END:VCALENDAR