Arabic speaker identification system using combination of DWT and LPC features

Speaker recognition plays a significant role in the field of human computer interaction. In the recent years, several researchers have contributed in this field and have successfully build machine learning models for automatic speaker recognition systems. In this paper, we propose an automatic speaker identification system for qaries (Quran reciter) of Arabic Language. For feature extraction discrete Wavelet Transform (DWT) and Linear Predictive Coding (LPC) feature extraction techniques were used. Classification was performed by Random Forest (RF). In order to improve the identification accuracy DWT and LPC features were used singly (One at a time) and combined to train RF. Our system showed the best performance when RF was trained with the combination of features. In this case 90.90% recognition accuracy was achieved.