High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

AUTOMATIC SPECTRAL ANALYSIS FOR SPEECH TO TEXT CONVERSION WITH EMBEDDED EXECUTION

Author(s):

P.Yasotha , Kongunadu college of Engineering; N.Rajasekaran, Kongunadu college of Engineering; M.Nathiya, Kongunadu college of Engineering

Keywords:

Speech Recognition, Mel Frequency Ceptral Coefficients, Hidden Markow Model, PIC Microcontroller, URAT Interface

Abstract

Speech signal important role in the Digital signal processing. In this paper speech sample observed with MFCC for the improvement of speech feature representation of the HMM based training approach. MFCC is used to extract the voice features from the voice sample and also HMM is used to the recognize the speaker based on the extracted features. I order to recognize the speaker first train the extracted feature by using HMM parameters and then compared to the original voice signal. If it is matched the voice signals to produce the text by using MATLAB Software. The Simulation results show an improvement of speech recognition with respect to computational time, learning precision for a speech recognition system. Here MATLAB output is interface to the Microcontroller by using UART and display through the LCD module.

Other Details

Paper ID: IJSRDV2I9371
Published in: Volume : 2, Issue : 9
Publication Date: 01/12/2014
Page(s): 644-646

Article Preview

Download Article