Number of the records: 1
Image Analysis by Methods of Dimension Reduction
- 1.0085056 - ÚI 2008 RIV US eng C - Conference Paper (international conference)
Moravec, P. - Snášel, V. - Frolov, A. - Húsek, Dušan - Řezanková, H. - Polyakov, P.Y.
Image Analysis by Methods of Dimension Reduction.
[Analýza obrazu pomocí metod redukce dimenze.]
Systems and Industrial Management Applications. Los Alamitos: IEEE Computer Society, 2007, s. 272-277. ISBN 0-7695-2894-5.
[CISIM'07. International Conference on Computer Information Systems and Industrial Management Applications /6./. Elk (PL), 28.06.2007-30.06.2007]
R&D Projects: GA ČR GA201/05/0079
Institutional research plan: CEZ:AV0Z10300504
Keywords : image analysis * methods of dimension reduction * cluster analysis
Subject RIV: BB - Applied Statistics, Operational Research
In this paper, we compare performance of several dimension reduction techniques, namely LSI, NMF, SDD, Boolean factor analysis (BFA), and cluster analysis. The qualitative comparison is evaluated on a collection of bars. The graphical comparison of the quality of these methods shows advantages of BFA.
Jsou porovnávány různé techniky redukce dimenze, zejména LSI, NMF, SDD, Boolovská faktorová analýza (BFA), a shluková analýza. Kvalitativní srovnání je provedeno na kolekci obrazu tvořených kombinací svislých a vodorovných čar (tzv. bars collection). Grafické porovnání ukazuje přednosti BFA.
Permanent Link: http://hdl.handle.net/11104/0147646
Number of the records: 1