Abstract—Speech recognition is progressively being utilized in practical applications with time. Automatic gender identification is one of the most intriguing applications since it distinguishes female and male speeches from briefly spoken communication records. This is advantageous in various applications, including automated conversation systems, system verification, demographic attribute prediction and assessing speaker’s expressions. Speech is a natural mode of communication, and pitch variation of a gender-specific speech signal is often used to identify a person as male or female. This paper presents a model to identify gender from Arabic speech by integrating audio preprocessing, Mel-Frequency Cepstral Coefficients (MFCC), Delta MFCC, and Log Filter bank feature extraction. Pre-processing involves testing pre-emphasis, framing, windowing, and Fast Fourier Transform. Finally, features are extracted using three feature extraction methods from the processed audios. Feed Forward Neural Networks and Keras-based Neural Networks are employed as classifier models. Regarding accuracy and simplicity, the proposed hybrid method surpasses most previous approaches discussed in the literature for gender categorization from Arabic speech. The proposed model achieved an average classification accuracy of 93.09%.
 
Keywords—speech recognition, Arabic language, gender classification, hybrid learning, technological development


Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.