|
|
Recognition of isolated words
Ondruška, Jiří ; Švrček, Martin (referee) ; Kolářová, Jana (advisor)
Human speech recognition in biometric systems is an actual problem, which science intensively deals with. One of most used methods is the method of hidden Markovov’s models. Attention in isolated words recognition is focused on characteristic speech signal parameters obtaining, enabling most clear identification due to hiddem Markov model application. This work concentrates on biometric systems, its methods, and then is focused on isolated words recognition problems. The hidden Markov model recognition system with usage of some Matlab functions is designed. Concept focuses on characteristic speech signal parameters obtaining, code book making through vector quantization, words model training and finally probability of concrete model and obtained word similarity evaluation. Ratio for one speaker's spoken words recognition reaches 40%.
|
|
|
Home as the topic of Czech philosophical work 20th and 21st century.
PŘEVRÁTILOVÁ, Pavlína
The diplom a thesis deals with home and its concepts in philosophical works. The aim was to grasp and describe home as a philosophical category, using the ideas and theories that we found in the Czech philosophical texts that came out during the 20th and 21st centuries. Through these available information, we have attempted to unify and create a more coherent, interconnected theory, in which we divided home intoseveral basic chapters that are absolutely decisive and decisive for the concept and understanding of home, both past and present.
|
|
|
Recognition of isolated words
Ondruška, Jiří ; Švrček, Martin (referee) ; Kolářová, Jana (advisor)
Human speech recognition in biometric systems is an actual problem, which science intensively deals with. One of most used methods is the method of hidden Markovov’s models. Attention in isolated words recognition is focused on characteristic speech signal parameters obtaining, enabling most clear identification due to hiddem Markov model application. This work concentrates on biometric systems, its methods, and then is focused on isolated words recognition problems. The hidden Markov model recognition system with usage of some Matlab functions is designed. Concept focuses on characteristic speech signal parameters obtaining, code book making through vector quantization, words model training and finally probability of concrete model and obtained word similarity evaluation. Ratio for one speaker's spoken words recognition reaches 40%.
|
guest ::
