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 325-AB\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221208T110000
DTEND;TZID=Asia/Seoul:20221208T123000
UID:siggraphasia_SIGGRAPH Asia 2022_sess168_papers_565@linklings.com
SUMMARY:LayoutEnhancer: Generating Good Indoor Layouts from Imperfect Data
DESCRIPTION:Technical Papers\n\nLayoutEnhancer: Generating Good Indoor Lay
 outs from Imperfect Data\n\nLeimer, Guerrero, Weiss, Musialski\n\nWe addre
 ss the problem of indoor layout synthesis, which is a topic of continuing 
 research interest in computer graphics. The newest works made significant 
 progress using data-driven generative methods; however, these approaches r
 ely on suitable datasets. In practice, some desirable layout properties ma
 y not exist in a dataset, for instance, specific expert knowledge can be m
 issing in the data.\n\nWe propose a method that combines expert knowledge,
  for example, knowledge about ergonomics, with a data-driven generator bas
 ed on the popular Transformer architecture. The knowledge is given as diff
 erentiable scalar functions, which can be used both as weights or as addit
 ional terms in the loss function. Using this knowledge, the synthesized la
 youts can be biased to exhibit desirable properties, even if these propert
 ies are not present in the dataset.\n\nOur approach can also alleviate pro
 blems of lack of data and imperfections in the data. Our work aims to impr
 ove generative machine learning for modeling and provide novel tools for d
 esigners and amateurs for the problem of interior layout creation.\n\nRegi
 stration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: ENGLISH\n\nF
 ormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_565&sess=sess16
 8
END:VEVENT
END:VCALENDAR