Thesis Details
Vícejazykový fonémový rozpoznávač
Aim, of this master thesis is training of phoneme recognizer with phoneme set, which have been made by merging of several phoneme sets, which are containted in SpeechDat-E database and find out if this kind of recognizer will have better results than recognizers which were trained on one language. This work also deals with phoneme sets, principles of phoneme recognition using recognizers based on artifical neural networks, language identification and merging of given phoneme sets. Also is described process of training phoneme recognizer and phoneme recognition.
Phonemes, IPA, International Phonetic Alphabet, SAMPA, LID, Language identification, PRML, PPRLM, HTK, STK, Phoneme recognizing, Merging of phoneme sets, Neural networks
Burget Radek, doc. Ing., Ph.D. (DIFS FIT BUT), člen
Kolář Dušan, doc. Dr. Ing. (DIFS FIT BUT), člen
Kotásek Zdeněk, doc. Ing., CSc. (DCSY FIT BUT), člen
Matoušek Petr, doc. Ing., Ph.D., M.A. (DIFS FIT BUT), člen
Racek Stanislav, doc. Ing., CSc. (WBU in Pilsen), člen
@mastersthesis{FITMT4950, author = "Vojt\v{e}ch Vobr", type = "Master's thesis", title = "V\'{i}cejazykov\'{y} fon\'{e}mov\'{y} rozpozn\'{a}va\v{c}", school = "Brno University of Technology, Faculty of Information Technology", year = 2007, location = "Brno, CZ", language = "czech", url = "https://www.fit.vut.cz/study/thesis/4950/" }