No abstract

Speech recognition system needs to perform a high complex calculation and short time to complete it. This is a big challenge for the real-time systems. However, using a simple and fast algorithm may do this task for the slow systems. Thus, the main objective of this paper is to design and implement a Real-Time Arabic Speech Recognition system using MATLAB environment. It is capable of accurately identifying some letters while remaining simple and fast. It uses the Mel-Frequency Cepstral Coefficients (MFCCs) as a feature extraction and Euclidean distance to compare the test sound and the database. A recognition rate of 89.6% has been reached.

No abstract