Phonetic recognition of Arabic figures based on formant analysis via artificial neural network
A. I. El Mallawany, M. I. Abdallah, M. Abd El Gawad, H. S. Sedik · The Journal of the Acoustical Society of America · 1999
This paper describes a speaker recognition for dependent and independent speaker systems based on the formant analysis technique. Also, this system is used for Arabic figure recognition. Linear predictive code (LPC) is applied for extracting the formants of the Arabic digits from 0–9. After training and selecting the proper network structure, the recognition accuracy is found to be 98.8% for the dependent system while it is 78.8% for the independent case. Recognition accuracy for independent speaker is improved to 95.5% when grouping the Arabic digits into three groups.