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:20240214T070247Z
LOCATION:Meeting Room C4.11\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231214T111000
DTEND;TZID=Australia/Melbourne:20231214T112500
UID:siggraphasia_SIGGRAPH Asia 2023_sess129_papers_431@linklings.com
SUMMARY:Diffusion Posterior Illumination for Ambiguity-aware Inverse Rende
 ring
DESCRIPTION:Technical Papers\n\nLinjie Lyu (Max-Planck-Institut für Inform
 atik), Ayush Tewari (MIT CSAIL), Marc Habermann (Max-Planck-Institut für I
 nformatik), Shunsuke Saito and Michael Zollhöfer (Reality Labs Research), 
 and Thomas Leimkühler and Christian Theobalt (Max-Planck-Institut für Info
 rmatik)\n\nInverse rendering, the process of inferring scene properties fr
 om images, is a challenging inverse problem. The task is ill-posed, as man
 y different scene configurations can give rise to the same image. Most exi
 sting solutions incorporate priors into the inverse-rendering pipeline to 
 encourage plausible solutions, but they do not consider the inherent ambig
 uities and the multi-modal distribution of possible decompositions. In thi
 s work, we propose a novel scheme that integrates a denoising diffusion pr
 obabilistic model pre-trained on natural illumination maps into an optimiz
 ation framework involving a differentiable path tracer. The proposed metho
 d allows sampling from combinations of illumination and spatially-varying 
 surface materials that are, both, natural and explain the image observatio
 ns. We further conduct an extensive comparative study of different priors 
 on illumination used in previous work on inverse rendering. Our method exc
 els in recovering materials and producing highly realistic and diverse env
 ironment map samples that faithfully explain the illumination of the input
  images.\n\nRegistration Category: Full Access\n\nSession Chair: Marc Stam
 minger (Friedrich-Alexander-Universität Erlangen-Nürnberg)
URL:https://asia.siggraph.org/2023/full-program?id=papers_431&sess=sess129
END:VEVENT
END:VCALENDAR