DESIGN AND IMPLEMENTATION OF VOICE RECOGNITION SYSTEM

ABSTRACT This project presents design and implementation of a voice recognition system that could identify different users, based on previously stored voice samples. Each user inputs audio samples with a keyword of his or her choice. This input is gathered, and successful processing to extract meaningful spectral coefficients will be achieved. These coefficients are to be stored in a database for later comparison with future audio inputs. Afterwards, the system has to capture an input from any other user and match its spectral coefficients to all previously stored coefficients on the database, in order to identify the unknown speaker. Since the spectral coefficients were acquired adequately the system as a whole did not recognize anything, although we believe that modifying the system’s structure to decrease timing dependencies between the subsystems might make the implementation more feasible and less complex.