Speech Recognition using MFCC and Neural Networks

  Divyesh S. Mistry,  Dr.A.V.Kulkarni

The most common mode of communication between humans is speech. As this is the most preferred way, humans would like to use speech to interact with machines also. That is why, automatic speech recognition has gained a lot of popularity. Many approaches for speech recognition exist like Dynamic Time Warping (DTW), Hidden Markov Model (HMM). This paper shows how Neural Network (NN) can be used for speech recognition and also investigates its performance in speech recognition. Learning Vector Quantization Neural Network has been applied. For the feature extraction of speech Mel Frequency Cepstrum Coefficients (MFCC) has been used which gives a set of feature vectors of speech waveform. Earlier research has shown MFCC to be more accurate and effective than other feature extraction techniques in the speech recognition. The work has been done on MATLAB and experimental results show that system is able to recognize words at sufficiently high accuracy.

Keywords- Speech Recognition ; Mel Frequency Cepstrum Coefficients (MFCC) ; Neural Networks ; Learning Vector Quantization

Unique Identification Number - IJEDR1402134
Page Number(s) - 2122-2129
Pubished in - Volume 2 | Issue 2 | June 2014
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Divyesh S. Mistry,  Dr.A.V.Kulkarni,   "Speech Recognition using MFCC and Neural Networks", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.2, Issue 2, pp.2122-2129, June 2014, Available at :http://www.ijedr.org/papers/IJEDR1402134.pdf