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:20260817T171531Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241203T165300
DTEND;TZID=Asia/Tokyo:20241203T170500
UID:siggraphasia_SIGGRAPH Asia 2024_sess110_papers_1057@linklings.com
SUMMARY:Efficient Neural Path Guiding with 4D Modeling
DESCRIPTION:Honghao Dong, Rui Su, Guoping Wang, and Sheng Li (Peking Unive
 rsity)\n\nPrevious local guiding methods used 3D data structures to model 
 spatial radiance variations but struggled with additional dimensions in th
 e path integral, such as temporal changes in dynamic scenes. Extending the
 se structures to higher dimensions also proves inefficient due to the curs
 e of dimensionality. In this study, we investigate the potential of compac
 t neural representations to model additional scene dimensions efficiently,
  thereby enhancing the performance of path guiding in specialized renderin
 g applications, such as distributed effects including motion blur. We pres
 ent an approach that models a higher dimensional spatio-temporal distribut
 ion through neural feature decomposition. Additionally, we present a cost-
 effective approximate with lower-dimensional representation to model only 
 subspace by progressive training strategy. We also investigate the benefit
 s of modeling correlations with the additional dimensions on typical distr
 ibuted ray tracing scenarios, including the motion blur effect in dynamic 
 scenes, as well as spectral rendering. Experimental results demonstrate th
 e effectiveness of our method on these applications.\n\nRegistration Categ
 ory: Full Access, Full Access Supporter\n\nLanguage Format: English Langua
 ge\n\nSession Chair: Michael Wimmer (TU Wien, Technische Universität Wien 
 (TU Wien))\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_1057&sess=sess110
END:VEVENT
END:VCALENDAR