BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163652Z
LOCATION:Meeting Room C4.8\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231215T093000
DTEND;TZID=Australia/Melbourne:20231215T094000
UID:siggraphasia_SIGGRAPH Asia 2023_sess155_papers_907@linklings.com
SUMMARY:Conditional Resampled Importance Sampling and ReSTIR
DESCRIPTION:Markus Kettunen and Daqi Lin (NVIDIA); Ravi Ramamoorthi (NVIDI
 A, University of California San Diego); Thomas Bashford-Rogers (University
  of Warwick); and Chris Wyman (NVIDIA)\n\nRecent work on generalized resam
 pled importance sampling (GRIS) enables importance-sampled Monte Carlo int
 egration with random variable weights replacing the usual division by prob
 ability density. This enables very flexible spatiotemporal sample reuse, e
 ven if neighboring samples (e.g., light paths) have intractable probabilit
 y densities.\nUnlike typical Monte Carlo integration, which samples accord
 ing to some PDF, GRIS instead resamples existing samples.  But resampling 
 with GRIS assumes samples have tractable marginal contribution weights, wh
 ich is problematic if reusing, for example, light subpaths from unidirecti
 onally-sampled paths.\nReusing such subpaths requires conditioning by (non
 -reused) segments of the path prefixes. \n\nIn this paper, we extend GRIS 
 to conditional probability spaces, showing correctness given certain condi
 tional independence between integration variables and their unbiased contr
 ibution weights.  We show proper conditioning when using GRIS over randomi
 zed conditional domains and how to formulate a joint unbiased contribution
  weight for unbiased integration.\n\nTo show our theory has practical impa
 ct, we prototype a modified ReSTIR PT with a final gather pass.  This reus
 es subpaths, postponing reuse at least one bounce along each light path.  
 As in photon mapping, such a final gather reduces blotchy artifacts from s
 ample correlation and reduced correlation improves the behavior of modern 
 denoisers on ReSTIR PT signals.\n\nRegistration Category: Full Access\n\nS
 ession Chair: Sing Chun Lee (Bucknell University)\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_907&sess=sess155
END:VEVENT
END:VCALENDAR