BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171533Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241203T153100
DTEND;TZID=Asia/Tokyo:20241203T154300
UID:siggraphasia_SIGGRAPH Asia 2024_sess107_papers_296@linklings.com
SUMMARY:NeRF-Casting: Improved View-Dependent Appearance with Consistent R
 eflections
DESCRIPTION:Dor Verbin, Pratul P. Srinivasan, Peter Hedman, and Ben Milden
 hall (Google Research); Benjamin Attal (Carnegie Mellon University); and R
 ichard Szeliski and Jonathan T. Barron (Google Research)\n\nNeural Radianc
 e Fields (NeRFs) typically struggle to reconstruct and render highly specu
 lar objects, whose appearance varies quickly with changes in viewpoint. Re
 cent works have improved NeRF's ability to render detailed specular appear
 ance of distant environment illumination, but are unable to synthesize con
 sistent reflections of closer content. Moreover, these techniques rely on 
 large computationally-expensive neural networks to model outgoing radiance
 , which severely limits optimization and rendering speed. We address these
  issues with an approach based on ray tracing: instead of querying an expe
 nsive neural network for the outgoing view-dependent radiance at points al
 ong each camera ray, our model casts reflection rays from these points and
  traces them through the NeRF representation to render feature vectors whi
 ch are decoded into color using a small inexpensive network. We demonstrat
 e that our model outperforms prior methods for view synthesis of scenes co
 ntaining shiny objects, and that it is the only existing NeRF method that 
 can synthesize photorealistic specular appearance and reflections in real-
 world scenes, while requiring comparable optimization time to current stat
 e-of-the-art view synthesis models.\n\nRegistration Category: Full Access,
  Full Access Supporter\n\nLanguage Format: English Language\n\nSession Cha
 ir: Hongzhi Wu (Zhejiang University; State Key Laboratory of CAD&CG, Zheji
 ang University)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_296&sess=sess107
END:VEVENT
END:VCALENDAR