Speech Defect Analysis Using Hidden Markov Models

Abstract

The main aim of this paper is the analysis of speech deteriorated by a very rare disease, which induce epileptic seizures in a part of brain responsible for speech production. Speech defects, represented mostly by the combination of missing and mismatched phonemes, are sought and examined in the spectral and time domain. An algorithm, proposed in this paper, is based on Hidden Markov Models (HMMs) and it is most suitable for the speech recognition tasks. The algorithm is able to analyze in both time and spectral domains simultaneously; in the spectral domain as a log-likelihood score and in the time domain as a forced time alignment of the HMMs. The suggested algorithm works properly in the time domain. The results for the spectral domain are not credible, because the algorithm have to be tested on more data (not available at the time of paper preparation).

Keywords

• Speech defects
• LKS
• developmental dysphasia
• HMMs
• speech recognition
• forced time alignment