Voice Activated E-Learning System for the Visually Impaired

E-learning has become an important tool for learners to acquire information and knowledge. However visually impaired people have no or very little access to this tool, since interface suitable to them are unavailable. The Voice Activated E learning System can provide a solution to this problem. Developing this system is meant to assist visually impaired students in learning, a desired subject, from the system, in a convenient way using their voice commands. This system consists of two major subsystems; namely Speaker Verification and Speech Recognition subsystem. In the speaker verification subsystem, Mel-Frequency Cepstral Coefficients (MFCC) is used for Feature extraction and Vector Quantization (VQ) algorithm is used for codebook generation. In the speech recognition subsystem, MFCC and dynamic programming (DP) are used. Experimental results show an accuracy of 96% in speaker verification subsystem and 89% in speech recognition subsystem.