Speech recognition using Gaussian mixture model with MFCC feature extraction
Loading...
Date
2020-09
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
This research work is about the design and analysis of the Speech Recognition Using Gaussian Mixture Model With MFCC Feature Extraction is implemented in the Linux operating system using python language. This thesis contains the basics and detailed view of speech recognition, in which different problems, applications and challenges are discussed. Speech recognition is related to biometric problem recognition, so pattern recognition is very important part. For the proposed model we have used Gaussian mixture model, it is a probabilistic model which work on expectation maximization algorithm. For feature extraction the proposed algorithm uses mel-frequency cepstral coefficient, which is replica of human hearing system. We extract features form each individual speaker audio and create a Gaussian mixture model for them. For prediction we give input speech signal, features are extracted from it then pattern is matched. The speaker having same pattern is the predicted speaker. Accuracy is measured with the help of confusion matrix in which for the proposed model we have calculated accuracy, precision, F1-score and recall.