Number of the records: 1
Transfer Learning of Mixture Texture Models
- 1.
SYSNO ASEP 0535433 Document Type C - Proceedings Paper (int. conf.) R&D Document Type Conference Paper Title Transfer Learning of Mixture Texture Models Author(s) Haindl, Michal (UTIA-B) RID, ORCID
Havlíček, Vojtěch (UTIA-B) RIDNumber of authors 2 Source Title Computational Collective Intelligence. - : Springer Nature Switzerland AG Cham, 2020 / Nguyen N. T. ; Hoang B. H. ; Huynh C. P. ; Hwang D. ; Trawinski B. ; Vossen G. - ISSN 0302-9743 - ISBN 978-3-030-63006-5 Pages s. 825-837 Number of pages 13 s. Publication form Print - P Action International Conference on Computational Collective Intelligence 2020 /12./ Event date 30.11.2020 - 03.12.2020 VEvent location Da Nang Country VN - Viet Nam Event type WRD Language eng - English Country CH - Switzerland Keywords Texture modeling ; transfer learning ; compound random field model ; bidirectional texture function Subject RIV BD - Theory of Information OECD category Automation and control systems R&D Projects GA19-12340S GA ČR - Czech Science Foundation (CSF) Institutional support UTIA-B - RVO:67985556 EID SCOPUS 85097519102 DOI 10.1007/978-3-030-63007-2_65 Annotation A transfer learning approach for multidimensional parametric mixture random field-based textural representation is introduced. The proposed transfer learning approach allows alleviating the multidimensional mixture models requirement for sufficiently large, but not always available, learning data sets. These compound random field models consist of an underlying structure model that controls transitions between several sub-models, each of them has different characteristics. The structure model proposed is a two-dimensional probabilistic mixture model, either of the Bernoulli or Gaussian mixture type. Local textures are modeled using the fully multispectral three-dimensional Gaussian mixture sub-models. Both presented compound random field models allow the reproduction of, compresses, edits, and enlarges a given measured color, multispectral, or bidirectional texture function (BTF) texture so that ideally, both measured and synthetic textures are visually indiscernible. Workplace Institute of Information Theory and Automation Contact Markéta Votavová, votavova@utia.cas.cz, Tel.: 266 052 201. Year of Publishing 2021
Number of the records: 1