CIVILICA We Respect the Science
(ناشر تخصصی کنفرانسهای کشور / شماره مجوز انتشارات از وزارت فرهنگ و ارشاد اسلامی: ۸۹۷۱)

گواهی نمایه سازی مقاله Bidirectional Neural Network for Pathological Voice Detection

عنوان مقاله: Bidirectional Neural Network for Pathological Voice Detection
شناسه (COI) مقاله: ICBME20_095
منتشر شده در بیستمین کنفرانس مهندسی زیست پزشکی ایران در سال ۱۳۹۲
مشخصات نویسندگان مقاله:

Iman Esmaili - Biomedical Engineering Department Science & Research Branch, Islamic Azad University Tehran, Iran
Nader Jafarnia Dabanloo - Biomedical Engineering Department Science & Research Branch, Islamic Azad University Tehran, Iran
keyvan Maghooli - Biomedical Engineering Department Science & Research Branch, Islamic Azad University Tehran, Iran

خلاصه مقاله:
We showed in our recent work that Bidirectional neural network (BNN) is a powerful tool for feature compensation in automatic speech recognition systems. In thispaper, we have introduced BNN as feature compensator for better discriminating of pathological voices from normal subjects. Mel-Frequency Cepstral Coefficients (MFCCs) wereextracted from each frame of sample voices and were compensated in two steps. First, BNN is trained with both normaland pathological feature vectors. Our hypothesis is that BNN can extract useful knowledge about the patterns of each class duringtraining step. In second step, MFCC feature vectors feed into BNN and compensate according to latent knowledge of BNN. In the last step , Compensated MFCCs are classified as pathological or normal by HMMs. We achieved 4.67%, 2.81% and 2.24% improvement in measures of specificity, accuracy and sensitivityby compensated feature vectors compared to the original feature vectors. Results corroborated our hypothesis about the ability ofBNN in compensation of feature vectors in a way that these features become more suitable for detection of pathological voices from normal ones.

کلمات کلیدی:
Bidirectional Neural Network, Pathological Voice Detection, Feature Compensation

صفحه اختصاصی مقاله و دریافت فایل کامل: https://www.civilica.com/Paper-ICBME20-ICBME20_095.html