Biometric Authentication: A Machine Learning Approach

1.1. Introduction

In business and personal life today, security protection systems are critical for many application domains: (1) transaction protection (client verification); (2) access control (key or keyless); (3) computer and network security (terminal user verification); and most important, (4) personal and public safety. Since almost all artificial features used in conventional identification techniques can potentially be forged and some of them (e.g., ID cards or passwords) stolen or forgotten,^[1] better and more effective identification and authentication methods are now in greater demand. With recent technological advances in audio and visual microelectronic systems, reliable automatic authentication systems have become a commercial ...