To search, Click below search items.


All Published Papers Search Service


A Novel Approach for Features Extraction Towards Classifying Normal and Special Children Speech Emotions in Urdu Language


Maria Andleeb, Najmi Ghani Haider, Saman Hina, Syed Abbas Ali


Vol. 17  No. 7  pp. 188-195


Spoken utterances play significant role in identifying the emotional states of speakers. However, extracted features add sense in spoken utterances which leads to provide speaker emotions. In this paper, a novel approach is presented for feature extraction toward classifying normal and special children speech emotion using spoken utterances in Urdu language. Eleven different features presents in this paper using thresholding technique on the extracted features implements the proposed algorithm namely: frequency, pitch, rate of zero passages, rate of acceleration, formant frequencies, intensity, log power, log energy, Mel Frequency Cepstrum Coefficients (MFCC), Linear Prediction Cepstral coefficient (LPCC) and Relative Spectral Transform - Perceptual Linear Prediction (Rasta PLP) with four different emotions (Angry, Happy, Neutral and Sad) to classify speech emotion of normal and special children. Experimental results evident that the proposed algorithm shows 100% accuracy with reduce error rate in Normal speech emotion and special speech emotion category for Angry, Neutral, Sad and Happy and Sad emotions respectively.


Feature Extraction, Speech Emotion, Normal and special children, Urdu Language, classification accuracy.