BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035312Z
LOCATION:Room 324\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221209T140000
DTEND;TZID=Asia/Seoul:20221209T153000
UID:siggraphasia_SIGGRAPH Asia 2022_sess174_papers_383@linklings.com
SUMMARY:Look-Ahead Training with Learned Reflectance Loss for Single-Image
  SVBRDF Estimation
DESCRIPTION:Technical Communications, Technical Papers\n\nLook-Ahead Train
 ing with Learned Reflectance Loss for Single-Image SVBRDF Estimation\n\nZh
 ou, Kalantari\n\nIn this paper, we propose a novel optimization-based meth
 od to estimate the reflectance properties of a near planar surface from a 
 single input image. Specifically, we perform test-time optimization by dir
 ectly updating the parameters of a neural network to minimize the test err
 or. Since single image SVBRDF estimation is highly ill-posed, such an opti
 mization is prone to\noverfitting. Our main contribution is to address thi
 s problem by introducing a training mechanism that takes the test-time opt
 imization into account. Specifically, we train our network by minimizing t
 he training loss after one or more gradient updates with the test loss. By
  training the network in this manner, we ensure that the network does not 
 overfit to the input image during the test-time optimization process. Addi
 tionally, we propose a learned reflectance loss to augment the typically u
 sed rendering loss during the test-time optimization. We do so by using an
  auxiliary network that estimates pseudo ground truth reflectance paramete
 rs and train it in combination with the main network. Our approach is able
  to converge with a small number of iterations of the test-time optimizati
 on and produces better results compared to the state-of-the-art methods.\n
 \nRegistration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: ENGLIS
 H\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_383&sess=sess17
 4
END:VEVENT
END:VCALENDAR