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:20260114T163641Z
LOCATION:Meeting Room C4.8\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231213T171600
DTEND;TZID=Australia/Melbourne:20231213T173100
UID:siggraphasia_SIGGRAPH Asia 2023_sess146_papers_463@linklings.com
SUMMARY:MetaLayer: A Meta-learned BSDF Model for Layered Materials
DESCRIPTION:Jie Guo, Zeru Li, and Xueyan He (Nanjing University); Beibei W
 ang (Nankai University, Nanjing University of Science and Technology); Wen
 bin Li and Yanwen Guo (Nanjing University); and Ling-Qi Yan (University of
  California, Santa Barbara)\n\nReproducing the appearance of arbitrary lay
 ered materials has long been a critical challenge in computer graphics, wi
 th regard to the demanding requirements of both physical accuracy and low 
 computation cost. Recent studies have demonstrated promising results by le
 arning-based representations that implicitly encode the appearance of comp
 lex (layered) materials by neural networks. However, existing generally-le
 arned models often struggle between strong representation ability and high
  runtime performance, and also lack physical/perceptual parameters for mat
 erial editing. To address these concerns, we introduce MetaLayer, a new me
 thodology leveraging meta-learning for modeling and rendering layered mate
 rials. MetaLayer contains two networks: a BSDFNet that compactly encodes t
 he appearance of layered materials, and a MetaNet that establishes the map
 ping between the physical parameters of each material and the weights of i
 ts corresponding implicit neural representation. A new positional encoding
  method and a well-designed training strategy are employed to improve the 
 performance and quality of the neural model. As a new learning-based repre
 sentation, the proposed MetaLayer model provides both fast responses to ma
 terial editing and high-quality results for a wide range of layered materi
 als, outperforming existing layered BSDF models.\n\nRegistration Category:
  Full Access\n\nSession Chair: Hongzhi Wu (Zhejiang University; State Key 
 Laboratory of CAD&CG, Zhejiang University)\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_463&sess=sess146
END:VEVENT
END:VCALENDAR