VOICE RECOGNITION SYSTEM: SPEECH-TO-TEXT

Abstract

VOICE RECOGNITION SYSTEM:SPEECH-TO-TEXT is a software that lets the user control computer functions and dictates text by voice. The system consists of two  components , first component is  for processing acoustic signal which is captured by a microphone and second component is to interpret the processed signal, then  mapping of the signal to words. Model for each letter will be built using Hidden Markov Model(HMM). Feature extraction will be done using Mel Frequency Cepstral Coefficients(MFCC). Feature training of the dataset will be done using vector quantization and Feature testing of the dataset will be done using viterbi algorithm. Home automation will be completely based on voice recognition system.