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:20231215T104500
DTEND;TZID=Australia/Melbourne:20231215T105500
UID:siggraphasia_SIGGRAPH Asia 2023_sess154_papers_1043@linklings.com
SUMMARY:ActRay: Online Active Ray Sampling for Radiance Fields
DESCRIPTION:Technical Papers\n\nJiangkai Wu, Liming Liu, Yunpeng Tan, Quan
 lu Jia, Haodan Zhang, and Xinggong Zhang (Peking University)\n\nThanks to 
 the high-quality reconstruction and photorealistic rendering, the Neural R
 adiance Field (NeRF) has garnered extensive attention and has been continu
 ously improved. Despite its high visual quality, the prohibitive training 
 time limits its practical application. Although significant acceleration h
 as been achieved, it is still far from real-time training, due to the need
  for tens of thousands of iterations. In this paper, a feasible solution i
 s to reduce the number of required iterations by always training the rays 
 with the highest loss values, instead of the traditional method of trainin
 g each ray with a uniform probability. To this end, we propose an online a
 ctive ray sampling strategy, ActRay. Specifically, to avoid the substantia
 l overhead of calculating the actual loss values for all rays in each iter
 ation, a rendering-gradient-based loss propagation algorithm is presented 
 to efficiently estimate the loss values. To further narrow the gap between
  the estimated loss and the actual loss, an online learning algorithm base
 d on the Upper Confidence Bound (UCB) is proposed to control the sampling 
 probability of the rays, thereby compensating for the bias in loss estimat
 ion. We evaluate ActRay on both real-world and synthetic scenes, and the p
 romising results show that it accelerates radiance field training to 6.5x.
  Besides, we test ActRay under all kinds of representations of radiance fi
 elds (implicit, explicit, and hybrid), demonstrating that it is general an
 d effective to different representations. We believe this work will contri
 bute to the practical application of radiance fields, because it has taken
  a step closer to real-time radiance field training.\n\nRegistration Categ
 ory: Full Access\n\nSession Chair: Yuchi Huo (Zhejiang University, Korea A
 dvanced Institute of Science and Technology)
URL:https://asia.siggraph.org/2023/full-program?id=papers_1043&sess=sess15
 4
END:VEVENT
END:VCALENDAR