Cleft palate speech initial and final automatic segmentation method and system

Abstract

The invention discloses a cleft palate speech initial and final automatic segmentation method and a cleft palate speech initial and final automatic segmentation system, which relate to the field of speech signal processing technologies, especially a cleft palate speech initial and final automatic segmentation technology. The invention aims to provide an initial and final automatic segmentation algorithm for cleft palate speech to solve the technical problem that an initial and final automatic segmentation algorithm for the cleft palate speech does not exist in the prior art. The cleft palate speech initial and final automatic segmentation method is characterized by carrying out the processing steps on a syllable including an initial: a quasi voiced initial and a quasi voiceless initial distinguish step, which is used for distinguishing whether the syllable is a syllable comprising a quasi voiced initial or one comprising a quasi voiceless initial; an initial and final segmentation step for the syllable comprising the quasi voiced initial; and an initial and final segmentation step for the syllable comprising the quasi voiceless initial.

Description

Technical field [0001] The invention relates to the technical field of speech signal processing, in particular to an automatic segmentation technology for the initials and finals of cleft palate speech. Background technique [0002] Cleft lip and palate is the most common congenital craniofacial deformity. Among them, speech and language dysfunction is the biggest impact of cleft palate. Speech assessment is an important part of the sequence therapy of cleft palate. It has important clinical significance for the realization of computer automatic evaluation algorithm for cleft palate speech. [0003] Dysarism is a unique manifestation of cleft palate speech. Dysarism is caused by incorrect pronunciation of the patient. Its manifestations include omission of consonants (consonants in syllables are omitted, and only vowels are left), substitution of consonants (a certain The consonant part of a syllable is replaced by another consonant, thus becoming a different syllable) and compe...

AI Technical Summary

Problems solved by technology

[0004] In the current research, there is an automatic segmentation of Mandarin consonants and finals, but these studies are based on the segmentation algorithm of Mandarin consonants and finals for normal people

At present, there are few researches on the segmentation algorithm of consonants and finals for pathological speech

Images