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:20240214T070242Z
LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T093000
DTEND;TZID=Australia/Melbourne:20231212T124500
UID:siggraphasia_SIGGRAPH Asia 2023_sess209_tog_109@linklings.com
SUMMARY:Spatiotemporally Consistent HDR Indoor Lighting Estimation
DESCRIPTION:Technical Papers\n\nZhengqin Li (Meta, University of Californi
 a San Diego); Yu Li and Mikhail Okunev (Meta); Manmohan Chandraker (Univer
 sity of California San Diego); and Zhao Dong (Meta)\n\nWe propose a physic
 ally-motivated deep learning framework to solve a general version of the c
 hallenging indoor lighting estimation problem. Given a single LDR image wi
 th a depth map, our method predicts spatially consistent lighting at any g
 iven image position. Particularly, when the input is an LDR video sequence
 , our framework not only progressively refines the lighting prediction as 
 it sees more regions, but also preserves temporal consistency by keeping t
 he refinement smooth. Our framework reconstructs a spherical Gaussian ligh
 ting volume (SGLV) through a tailored 3D encoder-decoder, which enables sp
 atially consistent lighting prediction through volume ray\ntracing, a hybr
 id blending network for detailed environment maps, an innetwork Monte-Carl
 o rendering layer to enhance photorealism for virtual object insertion, an
 d recurrent neural networks (RNN) to achieve temporally consistent lightin
 g prediction with a video sequence as the input. For training, we signific
 antly enhance the OpenRooms public dataset of photorealistic synthetic ind
 oor scenes with around 360K HDR environment maps of much higher resolution
  and 38K video sequences, rendered with GPU-based path tracing. Experiment
 s show that our framework achieves lighting prediction with higher quality
  compared to state-of-the-art single-image or video-based methods, leading
  to photorealistic AR applications such as object insertion.\n\nRegistrati
 on Category: Full Access, Enhanced Access, Trade Exhibitor, Experience Hal
 l Exhibitor
URL:https://asia.siggraph.org/2023/full-program?id=tog_109&sess=sess209
END:VEVENT
END:VCALENDAR