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:20230103T035311Z
LOCATION:Room 324\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221208T110000
DTEND;TZID=Asia/Seoul:20221208T123000
UID:siggraphasia_SIGGRAPH Asia 2022_sess164_papers_265@linklings.com
SUMMARY:Learning-Based Bending Stiffness Parameter Estimation by a Drape T
 ester
DESCRIPTION:Technical Papers\n\nLearning-Based Bending Stiffness Parameter
  Estimation by a Drape Tester\n\nFeng, Huang, Xu, Wang\n\nReal-world fabri
 cs often possess complicated nonlinear, anisotropic bending stiffness prop
 erties.  Measuring the physical parameters of such properties for physics-
 based simulation is difficult yet unnecessary, due to the persistent exist
 ence of numerical errors in simulation technology.  In this work, we propo
 se to adopt a simulation-in-the-loop strategy: instead of measuring the ph
 ysical parameters, we estimate the simulation parameters to minimize the d
 iscrepancy between reality and simulation.  This strategy offers good flex
 ibility in test setups, but the associated optimization problem is computa
 tionally expensive to solve by numerical methods.  Our solution is to trai
 n a regression-based neural network for inferring bending stiffness parame
 ters, directly from drape features captured in the real world.  Specifical
 ly, we choose the Cusick drape test method and treat multiple-view depth i
 mages as the feature vector.  To effectively and efficiently train our net
 work, we develop a highly expressive and physically validated bending stif
 fness model, and we use the traditional cantilever test to collect the par
 ameters of this model for 618 real-world fabrics.  Given the whole paramet
 er data set, we then construct a parameter subspace, generate new samples 
 within the subspace, and finally simulate and augment synthetic data for t
 raining purposes.  The experiment shows that our trained system can replac
 e cantilever tests for quick, reliable and effective estimation of simulat
 ion-ready parameters.  Thanks to the use of the system, our simulator can 
 now faithfully simulate bending effects comparable to those in the real wo
 rld.\n\nRegistration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: 
 ENGLISH\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_265&sess=sess16
 4
END:VEVENT
END:VCALENDAR