| Title |
Skaičių ir vardų garsynų atpažinimo, naudojant žodžiais grįstus paslėptuosius Markovo modelius, tyrimas / |
| Translation of Title |
Investigation of digits and names corpora recognition by words based hidden Markov models. |
| Authors |
Makauskas, Ignas |
| Full Text |
|
| Pages |
58 |
| Keywords [eng] |
hidden Markov Models ; HTK ; WEKA ; speech recognition |
| Abstract [eng] |
The final project presents research of numbers and names, spoken in Lithuanian, recognition system using hidden Markov models and Spanish speech recognizer. Possibilities to combine two different speech recognizers are also presented. In the methodological part preparation of speech recording database is described. Hidden Markov model toolkit and “WEKA” programs are analyzed. In research results part Lithuanian numbers and names recognition using hidden Markov models results are presented. They are obtained by changing number of states and Gaussian mixtures in hidden Markov models. Results of using Spanish speech identifier for the same words are also presented. Two different speech recognizers are combined into one system with decision making block for better recognizer selection. Different classifiers are used to control decision making block. The obtained results are compared and analyzed. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2017 |
