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:20240214T070249Z
LOCATION:Meeting Room C4.8\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231215T105500
DTEND;TZID=Australia/Melbourne:20231215T110500
UID:siggraphasia_SIGGRAPH Asia 2023_sess154_papers_729@linklings.com
SUMMARY:MCNeRF: Monte Carlo Rendering and Denoising for Real-Time NeRFs
DESCRIPTION:Technical Papers\n\nKunal Gupta (UC San Diego); Milos Hasan, Z
 exiang Xu, Fujun Luan, Kalyan Sunkavalli, and Xin Sun (Adobe Inc.); Manmoh
 an Chandraker (UC San Diego); and Sai Bi (Adobe Inc.)\n\nThe volume render
 ing step used in Neural Radiance Fields (NeRFs) produces highly photoreali
 stic results, but is inherently slow because it evaluates an MLP at a larg
 e number of sample points per ray. Previous work has addressed this by eit
 her proposing neural scene representations that are faster to evaluate or 
 by pre-computing (and approximating) scene properties to reduce render tim
 es. In this work, we propose \mcnerf, a \emph{general} Monte Carlo-based r
 endering algorithm that can speed up \emph{any} NeRF representation. We sh
 ow that the NeRF volume rendering integral can be efficiently computed via
  Monte Carlo integration using an importance sampling scheme based on ray 
 transmittance distributions. This allows us to, at render time, vary the n
 umber of color samples evaluated per ray to trade-off visual quality (nois
 e variance) against performance. These noisy Monte Carlo estimates can be 
 further denoised using an inexpensive image-space denoiser trained per-sce
 ne. We demonstrate that \mcnerf can be used to speed up NeRF representatio
 ns like TensoRF and Instant-NGP by $7\times$ while closely matching their 
 visual quality and without making the scene approximations that real-time 
 NeRF rendering methods usually make.\n\nRegistration Category: Full Access
 \n\nSession Chair: Yuchi Huo (Zhejiang University, Korea Advanced Institut
 e of Science and Technology)
URL:https://asia.siggraph.org/2023/full-program?id=papers_729&sess=sess154
END:VEVENT
END:VCALENDAR