Combined Feature Extraction Techniques and Naive Bayes Classifier for Speech Recognition

Abstract

Speech processing and consequent recognition are important areas of Digital Signal Processing since speech allows people to communicate more naturally and efficiently. In this work, a speech recognition system is developed for recognizing digits in Malayalam. For recognizing speech, features are to be extracted from speech and hence feature extraction method plays an important role in speech recognition. Here, front end processing for extracting the features is performed using two wavelet based methods namely Discrete Wavelet Transforms (DWT) and Wavelet Packet Decomposition (WPD). Naive Bayes classifier is used for classification purpose. After classification using Naive Bayes classifier, DWT produced a recognition accuracy of 83.5% and WPD produced an accuracy of 80.7%. This paper is intended to devise a new feature extraction method which produces improvements in the recognition accuracy. So, a new method called Discrete Wavelet Packet Decomposition (DWPD) is introduced which utilizes the hybrid features of both DWT and WPD. The performance of this new approach is evaluated and it produced an improved recognition accuracy of 86.2% along with Naive Bayes classifier.

Weighted Combination of Naive Bayes and LVQ Classifier for Fongbe Phoneme Classification

“…From their findings they argued that the bias peculiar to the naive Bayes rule is not really detrimental to phoneme classification performance. With S. Sunny et al present in [16], a new method called Discrete Wavelet Packet Decomposition (DWPD) has been introduced which utilizes the hybrid features of both Discrete Wavelet Transforms (DWT) and Wavelet Packet Decomposition (WPD). To show the performance of their method, they used Naive Bayes classifier.…”

Isolated Telugu Speech Recognition On T-DSCC And DNN Techniques

“…It is difficult to recognize the continuous speech. In this paper we are working on isolated word speech recognition technique [5].…”