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:20260114T163633Z
LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T093000
DTEND;TZID=Australia/Melbourne:20231212T124500
UID:siggraphasia_SIGGRAPH Asia 2023_sess209_papers_774@linklings.com
SUMMARY:Adaptive Shells for Efficient Neural Radiance Field Rendering
DESCRIPTION:Zian Wang and Tianchang Shen (NVIDIA, University of Toronto); 
 Merlin Nimier-David and Nicholas Sharp (NVIDIA); Jun Gao (NVIDIA, Universi
 ty of Toronto); Alexander Keller (NVIDIA); Sanja Fidler (NVIDIA, Universit
 y of Toronto); and Thomas Müller and Zan Gojcic (NVIDIA)\n\nNeural radianc
 e fields achieve unprecedented quality for novel view synthesis, but their
  volumetric formulation remains expensive, requiring a huge number of samp
 les to render high-resolution images. Volumetric encodings are essential t
 o represent fuzzy geometry such as foliage and hair, and they are well-sui
 ted for stochastic optimization. Yet, many scenes ultimately consist large
 ly of solid surfaces which can be accurately rendered by a single sample p
 er pixel. Based on this insight, we propose a neural radiance formulation 
 that smoothly transitions between volumetric- and surface-based rendering,
  greatly accelerating rendering speed and even improving visual fidelity. 
 Our method constructs an explicit mesh envelope which spatially bounds a n
 eural volumetric representation. In solid regions, the envelope nearly con
 verges to a surface and can often be rendered with a single sample. To thi
 s end, we generalize the NeuS formulation with a learned spatially-varying
  kernel size which encodes the spread of the density, fitting a wide kerne
 l to volume-like regions and a tight kernel to surface-like regions. We th
 en extract an explicit mesh of a narrow band around the surface, with widt
 h determined by the kernel size, and fine-tune the radiance field within t
 his band. At inference time, we cast rays against the mesh and evaluate th
 e radiance field only within the enclosed region, greatly reducing the num
 ber of samples required. Experiments show that our approach enables effici
 ent rendering at very high fidelity. We also demonstrate that the extracte
 d envelope enables downstream applications such as animation and simulatio
 n.\n\nRegistration Category: Full Access, Enhanced Access, Trade Exhibitor
 , Experience Hall Exhibitor\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_774&sess=sess209
END:VEVENT
END:VCALENDAR