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:20250110T023313Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T093400
DTEND;TZID=Asia/Tokyo:20241206T094600
UID:siggraphasia_SIGGRAPH Asia 2024_sess140_papers_717@linklings.com
SUMMARY:Differentiable Photon Mapping using Generalized Path Gradients
DESCRIPTION:Technical Papers\n\nJiankai Xing and Zengyu Li (Tsinghua Unive
 rsity), Fujun Luan (Adobe Research), and Kun Xu (Tsinghua University)\n\nP
 hoton mapping is a fundamental and practical Monte Carlo rendering techniq
 ue for efficiently simulating global illumination effects, especially for 
 caustics and specular-diffuse-specular (SDS) paths. In this paper, we pres
 ent the first differentiable rendering method for photon mapping. The core
  of our method is a newly introduced concept named generalized path gradie
 nts. Based on the extended path space manifolds (EPSMs), the generalized p
 ath gradients define the derivatives of the vertex positions and color con
 tributions of a path with respect to scene parameters under given geometri
 c constraints. By formalizing photon mapping as a path sampling technique 
 through vertex merging and incorporating a smooth differentiable density e
 stimation kernel, we enable the differentiation of the photon mapping algo
 rithms based on the theoretical results of generalized path gradients. Exp
 eriments demonstrate that our method is more effective than state-of-the-a
 rt physics-based differentiable rendering methods in inverse rendering app
 lications in handling difficult illumination paths, especially SDS paths.\
 n\nRegistration Category: Full Access, Full Access Supporter\n\nLanguage F
 ormat: English Language\n\nSession Chair: Seungyong Lee (POSTECH)
URL:https://asia.siggraph.org/2024/program/?id=papers_717&sess=sess140
END:VEVENT
END:VCALENDAR