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 S426+S427\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251215T134200
DTEND;TZID=Asia/Hong_Kong:20251215T135300
UID:siggraphasia_SIGGRAPH Asia 2025_sess106_papers_1809@linklings.com
SUMMARY:Sample Space Partitioning and Spatiotemporal Resampling for Specul
 ar Manifold Sampling
DESCRIPTION:Pengpei Hong (University of Utah), Meng Duan (Nankai Universit
 y), Beibei Wang (Nanjing University), Cem Yuksel (University of Utah), and
  Tizian Zeltner and Daqi Lin (NVIDIA)\n\nCaustics rendering remains a long
 -standing challenge in Monte Carlo rendering because high-energy specular 
 paths occupy only a small region of path space, making them difficult to s
 ample effectively. Recent work such as Specular Manifold Sampling (SMS) [Z
 eltner et al. 2020] can stochastically sample these specular paths and est
 imate their unbiased weights using Bernoulli trials. However, applying SMS
  in interactive rendering is non-trivial because it is slow and delivers n
 oisy images given a very limited time budget.\n\nIn this work, we extend S
 MS for high-quality caustic rendering in interactive settings using sample
  space partitioning.  Our insight is that Newton iterations, the main perf
 ormance bottleneck of SMS, can be restricted to the vicinity of the seed p
 ath, which can dramatically improve the performance. We achieve this with 
 tile-based sample space partitioning, which bounds the manifold walk regio
 n and allows building a per-frame prior distribution that concentrates ini
 tial guesses around solutions. This reduces the cost of SMS and improves i
 ts sampling quality. Applying spatiotemporal reuse (ReSTIR) further amorti
 zes the sample generation cost, greatly increasing the effective sample co
 unt. As a result, we achieve significant variance reduction compared to SM
 S in interactive rendering scenarios.\n\nRegistration Category: Full Acces
 s, Full Access Supporter\n\nSession Chair: Yuchi Huo (Zhejiang University,
  Korea Advanced Institute of Science and Technology)\n\n
END:VEVENT
END:VCALENDAR