Automatic speech segmentation using the Arabic phonetic database

In this paper, a special algorithm is developed to accomplish segmentation for the Arabic speech based on energy level of the uttered words or sentences. In Arabic language, phonemes can be divided into two energy regions: unvoiced phonemes are categorized as low energy, for example, the sounds / ??? / (/s/) and / ??? / (/h/); On the other hand, vowels and semi-vowels such as / ??? / (???) (/ 'a /) and / ??? / (/w/) are categorized as high energy. Also included in this category are voiced fricatives; for instance the sounds / ??? / (/z/) and / ??? / (/aa/). The Arabic Phonetic Database (KAPD) developed by the Computer and Electronics Research Institute at KACST is utilized to perform the automatic segmentation of speech and test the developed algorithm.

[1]  Kris Demuynck,et al.  A Comparison of Different Approaches to Automatic Speech Segmentation , 2002, TSD.

[2]  Ossama Essa.  Using prosody in automatic segmentation of speech , 1998, ACM-SE 36.

[3]  Lamia Bouafif,et al.  Pitch detection and formant analysis of Arabic speech processing , 2001 .

[4]  Diego H. Milone,et al.  Evolutionary algorithm for speech segmentation , 2002, Proceedings of the 2002 Congress on Evolutionary Computation. CEC'02 (Cat. No.02TH8600).

[5]  Jan P. van Hemert,et al.  Automatic segmentation of speech , 1991, IEEE Trans. Signal Process..

[6]  T. M. Nazmy,et al.  A Novel Method for Arabic Consonant/Vowel Segmentation Using Wavelet Transform , 2005, Egypt. Comput. Sci. J..

[7]  Ossama Emam,et al.  Language Model Based Arabic Word Segmentation , 2003, ACL.

[8]  Manish Sharma,et al.  Subword-based text-dependent speaker verification system with user-selectable passwords , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.