Syllable based ASR system for isolated Tamil words using fuzzy based neural network

Authors
Kiruthika Krishnamoorthy 1
Rathinavelu Arumugam 2
1 Department of Computer Science and Engineering, Dr. Mahalingam College of Engineering and Technology, Pollachi
2 Department of Computer Science and Engineering, Dr. Mahalingam College of Engineering and Technology, Pollachi.
10.22034/jcse.2025.182114
Abstract
Speech recognition, also known as Automatic Speech recognition or computer speech recognition means
understanding the voice and performing any required task or the ability to match a voice against a provided or
acquired vocabulary. Automatic speech recognition (ASR) is an important topic of speech processing. Tamil words are
analyzed and recognized with respect to syllables and features are to be extracted. In this research, we present a
new approach to automatically segment the isolated Tamil words into syllable-like segments. The algorithm for
syllable segmentation works by processing the short-term energy function of the continuous speech signal. The
proposed approach for speech recognition uses the combined features of Mel-Frequency Cepstral Coefficients
(MFCC) and Linear Predictive Coding (LPC). Recognition process uses Fuzzy based Neural Network is combined
together with Feed Forward Neural Network to recognize the corresponding word.