0410563 20240103182222.720060210235959.9 3-211-83651-9 Wien Springer 2001 4 s. 406-409KůrkováJ.NerudaR.KárnýM.SteeleN. C. pattern recognition Gaussian mixture model kernel density estimate cav_un_auth*0212668 Paclík P. NL cav_un_auth*0101171 Novovičová Jana UTIA-B Ústav teorie informace a automatizace AV ČR, v. v. i. 12B 09K VS96063 GA MŠk cav_un_auth*0025066 KSK1075601 GA AV ČR cav_un_auth*0027435 AV0Z1075907 The method for complete mixture initialization based on a product kernel estimate of probability density function is proposed for mixture estimation using EM-algorithm. The mixture components are assumed to correspond to local maxima of optimaly smoothed kernel density estimate. The gradient method is used for local extrema finding. As the last step, agglomerative hiearchical clustering methods merges closest components together. A comparison to scale-space approaches is given on examples. cav_un_auth*0212760 International Conference on Artificial Neural Nets and Genetic Algorithms /5./ Prague CZ 22.04.2001-25.04.2001 BB RO http://hdl.handle.net/11104/0130652 UTIA-B 20010032 2001 2001 Wien Springer 3-211-83651-9 Artificial Neural Nets and Genetic Algorithms. Proceedings 406 409 Kůrková J. 340 Neruda R. 340 Kárný M. 340 Steele N. C. 340