HMM based Automatic Speech Recognition Analysis

This project's 'HMM Based Automatic Speech Recognition Analysis main motive is just to generate an Automatic speech recognition which is clear an accurate using Hidden Markov Model (HMM) to get accurate results at number of frequency ranges related to human voice. Here is a record of 12 different words which is recorded by using a number of different speakers that includes male and female both (especially female). Thereafter, the speech recognitions result reports are compared with different feature extraction methods in this project instead of one method. Because, an Earlier research work on this project thesis only one feature extraction method has been used and also using a recognition of seven small vocal sounds using HMM (Hidden Markov Model). This speech recognition system mainly divided into two major blocks here in this project. First Block includes the recording data base and feature extraction of all recorded signals. Here we use Mel frequency cepstral coefficients (MFCC), linear cepstral coefficients and fundamental frequency as feature extraction methods instead of one extraction method which were used earlier. For obtaining a Mel frequency cepstral coefficients (MFCC), a signal is passing through following parameters named as pre emphasis, framing, applying window function, Fast Fourier transform, filter bank and then discrete cosine transform, where as a linear frequency cepstral coefficients does not use Mel frequency. Now the Second part includes the description of HMM used for modeling and recognizing the spoken words. All the raining samples are clustered using K-means algorithm. Gaussian mixture containing mean, variance and weight are modeling parameters. Here is also a role of Baum Welch algorithm. it is used for training the samples and reestimate the parameters. Finally in the thesis, the Viterbi algorithm recognizes best sequence that exactly matches for given sequence there is given during speech vocal sounds which has to be recognized. Here all the simulations are done by using the MATLAB tool. Keywords---MATLAB, Rule Viewer, Operating System window 7, HMM, MFCC, Window Techniques, Feature Extraction Methods.