Course: Speech Analysis and Recognition

« Back
Course title Speech Analysis and Recognition
Course code KKY/ARŘ
Organizational form of instruction Lecture + Tutorial
Level of course Master
Year of study not specified
Semester Winter
Number of ECTS credits 6
Language of instruction Czech
Status of course Compulsory-optional, Optional
Form of instruction Face-to-face
Work placements This is not an internship
Recommended optional programme components None
Lecturer(s)
  • Pražák Aleš, Ing. Ph.D.
  • Psutka Josef, doc. Ing. Mgr. Ph.D.
  • Psutka Josef, Prof. Ing. CSc.
Course content
Introduction into processing, training and speech recognition. Representation of speech, speech production, speech units, model of speech production, hearing model, phonetic transcription, speech signal analysis, transmission channel, speech compression, LPC, MFCC, PLP, phonetic analysis, statistical approach to speech recognition, HMM, language modeling, decoding techniques, MAP, Viterbi, stack decoder, time-synchronous search, transducers, N best hypotheses.

Learning activities and teaching methods
Lecture supplemented with a discussion, One-to-One tutorial, Individual study, Lecture with visual aids, Practicum
  • Contact hours - 26 hours per semester
  • Practical training (number of hours) - 39 hours per semester
  • Individual project (40) - 40 hours per semester
  • Presentation preparation (report) (1-10) - 5 hours per semester
  • Preparation for an examination (30-60) - 45 hours per semester
prerequisite
Knowledge
orientovat se v teorii pravděpodobnosti a statistice
aplikovat základy lineární algebry
Skills
algoritmizovat
dekomponovat úlohu na subúlohy
aplikovat základy teorie pravděpodobnosti
Competences
N/A
learning outcomes
Knowledge
diskutovat teoretická východiska při návrhu parametrizačních technik řeči
vysvětlit základy statistického rozpoznávání řeči
formulovat princip HMM
orientovat se v jazykových modelech
formulovat problematiku dekódování v systémech rozpoznávání řeči
Skills
prakticky realizovat parametrizaci řečového signálu
natrénovat jednoduchý akustický HMM model
vytvořit jednoduchý jazykový model
realizovat jednoduché dekódovací strategie
Competences
N/A
teaching methods
Knowledge
Lecture with visual aids
One-to-One tutorial
Skills
Task-based study method
Individual study
Competences
Lecture with visual aids
Task-based study method
assessment methods
Knowledge
Combined exam
Skills
Seminar work
Competences
Combined exam
Seminar work
Recommended literature
  • Psutka, Josef. Mluvíme s počítačem česky. Praha : Academia, 2006. ISBN 80-200-1309-1.


Study plans that include the course
Faculty Study plan (Version) Category of Branch/Specialization Recommended year of study Recommended semester