Loading...

Pathology Analysis and Multi-Class Discrimination for Laryngeal Disorders

Pakravan, Mansooreh | 2012

586 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 43311 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Jahed, Mehran
  7. Abstract:
  8. Ability to speak lucidly plays a key role in social relations. Consequently the role of larynx is quite important and timely diagnosis of laryngeal diseases has proved to be crucial. Since conventional diagnostic methods of the larynx are usually expensive or bothersome, the aim of this project is to analyze and classify diseases of the larynx with the aid of signal processing which tend to be faster and easier to implement and quite economical. This study utilizes the vowel sound /a/ and a well referenced database, namely MEEI (Massachusetts Eye and Ear Infirmary) which includes 53 normal and 213 abnormal voices in 7 classified diseases. In this work, using existing signal modeling and processing methods, appropriate glottal waveform was obtained for each voice. After extraction of seven discriminating features and upon change of the feature space using Kernel Principle Component Analysis (KPCA), the classifier which is consisted of two classifiers (Naïve Bayes and Fisher Linear Discriminant) is exerted on the feature sets. Regarding voice pathology detection, the proposed approach achieved a classification accuracy of 95.83% for normal and Polyp classification, 95% for Edema and Paralysis classification, 88.98% for three class classification (Adductor, Paralysis and Nodules), and 79.93% for four class classification (Adductor, Polyp, Nodules and Paresis). As a result it can be concluded that the proposed method is comparatively an appropriate noninvasive approach that is faster, easier and most economical.

  9. Keywords:
  10. Speech Signal ; Feature Extraction ; Speech Processing ; Laryngeal Disorder ; Glottal Waveform ; Multi-Class Discrimination ; Kernel Principle Component Analysis (KPCA) ; Fisher Linear Discriminant

 Digital Object List

 Bookmark

No TOC