Mobile phone speaker recognition application

• Main Author: Wong, Joseph Pin Jie
• Other Authors: Chng Eng Siong
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Pattern recognition; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
• Online Access: http://hdl.handle.net/10356/63062
• Institution: Nanyang Technological University

Smartphones are being more integrated with our lives. Hence, security and privacy of the personal data stored in the smartphone is an increasing concern. To combat this issue, one solution is to implement a biometric system, such as a speaker recognition system, that would provide a higher level of verification of the user than the default password or pattern locks. The purpose of this project is to implement a speaker recognition system on an Android smartphone, namely a Samsung Galaxy S5. The system will allow users to enrol themselves on the Android Application, which will store their speaker model onto a web server. When the user tries to unlock their phone thereafter, the application will request that the user verifies themselves. Verification will happen on the server side and the result will be transmitted back to the user through the application. Experiments were done on various speakers to test the accuracy of the system. In the future, as more speaker recognition techniques are developed, new experiments can be done. In addition, other biometrics, such as facial recognition or thumbprint, could also be added to the system to improve the security of the system.