A Pathological Multi-Vowels Recognition Algorithm Based on LSP Feature

Abstract

At present, pathological voice recognition is mainly based on the classification of pathological voice. However, almost all the researches are based on the single vowel \a\ samples, but few on multivowels. In addition, the current researches on multi-vowels recognition are mainly for normal voices, which are unsuitable for the speech recognition of normal and pathological multi-vowels simultaneously. This paper concentrates on developing an accurate and robust feature called enhanced-bark line spectrum pair (E-BLSP) to detect and classify normal and pathological multi-vowels. We explore the impact of E-BLSP feature on recognition performance and propose an effective method based on the combination of three features including E-BLSP for pathological and normal multi-vowels. In this paper, first LSP and difference of adjacent LSP (DAL) features of a vowel are extracted. Then LSP feature is warped at bark domain to get bark line spectrum pair (BLSP). In addition, then E-BLSP feature is calculated by adjusting BLSP using DAL feature. Finally, the adjusted E-BLSP feature and other two traditional features, including linear prediction cepstrum coefficient (LPCC) and mel-frequency cepstrum coefficients (MFCC) are applied to support vector machine (SVM) and deep neural network (DNN) classifiers to explore the classification performance of single feature and feature combinations for pathological and normal vowels /a/, /i/ and /u/. The results show that the highest achieved accuracies for DNN and SVM network are 98.6190% and 96.2693%, while the largest achieved area under curves (AUC) are 0.9925 and 0.9868, correspondingly with the combination of three features including LPCC, MFCC, and E-BLSP.

Matching Pursuit and Sparse Coding for Auditory Representation

“…Speech signals in the time domain are transformed into the time-frequency domain because changes in frequencies and in the time of the signals are important cues for further analysis processes. On the basis of spectrograms, feature-emphasis techniques such as mel-frequency cepstral coefficients (MFCCs) [2]- [4] and linear prediction cepstral coefficients (LPCCs) [5], [6] can be applied to obtain significant acoustical feature vectors. Pattern-analysis techniques (e.g., support vector machine [7] for voice pathology detection, Gaussian mixture model [8] for speaker recognition, and deep neural networks [9] for emotion recognition) are then applied to the feature vectors.…”