Recognition method of cleft palate speech nasal leakage based on recursive graph analysis

Abstract

The invention discloses a recognition method of cleft palate speech nasal leakage based on recursive graph analysis, which relates to the field of speech signal processing. The detection method comprises the following steps: (1) preprocessing the speech signal, downsampling the input speech signal, and performing normalizing framing, pre-emphasis and amplitude normalization; (2) obtaining a recursive graph and a matrix recursive graph of the speech signal after the preprocessing; (3) performing trend analysis of recursive graph; (4) directly dividing the region into blocks for recursive graphand performing matrix processing for recursive graph; (5) performing image processing on the recursive image, and subjecting the image matrix after format conversion to binarization processing and filtering with two special structure templates in turn; (6) classifying and recognizing the signal by using classifier to obtain an automatic recognition result. Compared with the prior art, the detection result is objective and accurate, realizes a higher degree of automatic measurement, provides reliable reference data for the clinical digital evaluation of cleft palate speech nasal leakage, meetsthe development needs of precision medicine, and performs more accurate and effective signal classification and recognition.

Description

technical field [0001] The invention relates to the field of speech signal processing, in particular to a recognition method for cleft palate speech nasal air leakage based on recursive graph analysis. Background technique [0002] One of the most obvious functional impairments of congenital cleft palate deformity is speech impairment. Because the cleft palate loses the normal bony and muscular separation between the mouth and nose, patients cannot control the size and direction of airflow during pronunciation. Due to velopharyngeal insufficiency in patients with cleft palate, the velopharyngeal valve cannot be completely closed during pronunciation, and the airflow passes through the nasal cavity and oral cavity at the same time, resulting in typical cleft palate speech such as abnormal resonance, nasal air leakage, and high nasal sound. When the speaker pronounces consonants, there is airflow leaking from the nasal cavity, and the sound of this airflow passing through the ...

AI Technical Summary

Problems solved by technology

[0003] In recent years, scholars at home and abroad have devoted themselves to the research of more objective digital speech processing and recognition methods, but there are few researches on the diagnosis and evaluation of rhinorrhea speech in the prior art, and there is a lack of effective automatic recognition of rhinorrhea speech. Identification detection

Images