Speech Recognition using Artificial Neural Network

Humans prefer to interact with each other using speech. Since this is the most natural mode of communication, the humans also want to interact with machines using speech only. So, automatic speech recognition has gained a lot of popularity. Different approaches for speech recognition exists like Hidden Markov Model (HMM), Dynamic Time Warping (DTW), Vector Quantization (VQ), etc. This paper uses Neural Network (NN) along with Mel Frequency Cepstrum Coefficients (MFCC) for speech recognition. Mel Frequency Cepstrum Coefiicients (MFCC) has been used for the feature extraction of speech. This gives the feature of the waveform. For pattern matching FeedForward Neural Network with Back propagation algorithm has been applied. The paper analyzes the various training algorithms present for training the Neural Network and uses train scg for the experiment. The work has been done on MATLAB and experimental results show that system is able to recognize words at sufficiently high accuracy.