Acoustic/Phonetic Transcription Using a Polynomial Classifier and Hidden Markov Models

Authors

  • Alfred Kaltenmeier

Abstract

This paper describes a module for acoustic/phonetic transcription in a continuous speech understanding system. This module segments input utterances into sequences of phone classes which belong to one of six broad phonetic categories. In a higher system level such segment sequences are used to hypothesize possible word candidates from a lexicon.

This module is hierarchically implemented in two stages: a polynomial classifier for a frame-by-frame classification of phone classes followed by a segmentation stage using Hidden Markov Models (HMM) of phone class segments.

Additional Files

Published

2022-12-03

How to Cite

1.
Kaltenmeier A. Acoustic/Phonetic Transcription Using a Polynomial Classifier and Hidden Markov Models. Canadian Acoustics [Internet]. 2022 Dec. 3 [cited 2024 Nov. 8];14(3 bis):95-6. Available from: https://jcaa.caa-aca.ca/index.php/jcaa/article/view/3537

Issue

Section

Proceedings of the Acoustics Week in Canada