Abstract
The aim of this project is to produce a speaker voice recognition system based on a neural network approach. The voice sample is a one word pronounciation which is obtained from every speaker. Basically, voice patterns are different for each speaker. The neural network approach is used to obtain a single optimized value to represent the speaker voice pattern. The main focus of this project is to develop a voice recognition system in which the identity of the speaker is verified. In the recognition process, the voice pattern of each speaker is compared to the unknown voice pattern. The one that is closed to the unknown pattern is picked up for the verification.
Website : http://ir.fsksm.utm.my/2116/
No comments:
Post a Comment