Thesis Details

Adaptace rozpoznávače řeči na datech bez přepisu

Master's Thesis Student: Švec Ján Academic Year: 2014/2015 Supervisor: Schwarz Petr, Ing., Ph.D.
English title
Unsupervised Adaptation of Speech Recognizer
Language
Czech
Abstract

The goal of this thesis is to design and test techniques for unsupervised adaptation of speech recognizers on some audio data without any textual transcripts. A training set is prepared at first, and a baseline speech recognition system is trained. This sistem is used to transcribe some unseen data. We will experiment with an adaptation data selection process based on some speech transcript quality measurement. The system is re-trained on this new set than, and the accuracy is evaluated. Then we experiment with the amount of adaptation data.

Keywords

speech recognition, acustic model, language model, confidence, adaptation

Department
Department of Computer Graphics and Multimedia FIT BUT
Degree Programme
Information Technology, Field of Study Computer Graphics and Multimedia
Files
Status
defended, grade E
Date
25 June 2015
Reviewer
Karafiát Martin, Ing., Ph.D.
Committee
Smrž Pavel, doc. RNDr., Ph.D. (DCGM FIT BUT), předseda
Herout Adam, prof. Ing., Ph.D. (DCGM FIT BUT), člen
Rogalewicz Adam, doc. Mgr., Ph.D. (DITS FIT BUT), člen
Sedlák Petr, doc. Ing., Ph.D. (DPHYS FEEC BUT), člen
Szőke Igor, Ing., Ph.D. (DCGM FIT BUT), člen
Vašíček Zdeněk, doc. Ing., Ph.D. (DCSY FIT BUT), člen
Citation
ŠVEC, Ján. Adaptace rozpoznávače řeči na datech bez přepisu. Brno, 2015. Master's Thesis. Brno University of Technology, Faculty of Information Technology. 2015-06-25. Supervised by Schwarz Petr. Available from: https://www.fit.vut.cz/study/thesis/17301/
BibTeX 
@mastersthesis{FITMT17301,
    author = "J\'{a}n \v{S}vec",
    type = "Master's thesis",
    title = "Adaptace rozpozn\'{a}va\v{c}e \v{r}e\v{c}i na datech bez p\v{r}episu",
    school = "Brno University of Technology, Faculty of Information Technology",
    year = 2015,
    location = "Brno, CZ",
    language = "czech",
    url = "https://www.fit.vut.cz/study/thesis/17301/"
}

Theses

Back to top