Date of Publication :7th November 2016

Abstract: The primary purpose of this research work is to develop a speech recognition system, which will be capable of recognizing and interpreting an individual’s voice. This will thus obviate the use of keyboards and other such input devices to output the data. The scope of this work is limited to the recognition of digits. Matlab, a versatile software tool is made use of for the processing of the signals. This processing involves filtering, shaping, correlating the speech signal which will enable the computer to interpret the commands given by the user. The signature of every spoken digit is unique from which we can derive the parameters which differ from digit to digit. These parameters serve as a concrete base for the implementation of a speaker independent voice recognition system. Using such parameters simultaneously and using probability concept the digits are sorted and arranged in the descending order of their probability. The digit with the highest probability is displayed.

Reference :