Joint Conferences Tutorial

Reliable identification of humans is very important in a number of applications including law enforcement (e.g., locating missing children), government (e.g., border control), commercial (e.g., web access) applications, and other access control and e-commerce applications. Current approaches for identity recognition are based on pass cards or PIN numbers that can be stolen or forgotten and fail to provide non-reputability at the transaction level. Biometrics-based authentication techniques utilize physiological and behavioral characteristics (e.g., fingerprint, iris, and face) to either verify a claimed identity (biometric verification) or establish the identity of an individual (biometric identification). With increased emphasis on security, there is a growing and urgent need to identify humans both locally and remotely on a routine basis. Over the past decade, significant advances have been made in algorithms for biometric recognition.

The goal of this tutorial is to discuss the different biometric recognition algorithms (e.g., face, voice, fingerprint and iris, etc.) and to understand the relative advantages and limitations of different approaches. The tutorial will summarize the state of the art methods and results in biometric recognition. We will also discuss the methods based on multiple biometrics (multi-modal biometric) recognition and show the improved results that could be obtained. Issue related to requirements for different biometrics, different ways to choose o biometrics, techniques used to compare and test different biometrics will also be discussed. MATLAB demonstrations will be used to illustrate the different concepts.

The tutorial is designed for engineers, students, researchers, and managers who want to learn the fundamentals behind biometric technology. The course is of particular interest to engineers who are members of projects that involve analysis and implementation of biometric systems. Students wanting to pursue advanced studies and/or careers in this area will benefit greatly from the tutorial. Some prior background in probability theory, some familiarity with vectors and matrices, and basic signals and systems theory is needed.

Mohamed Deriche received his BS in Electrical Engineering from the National Polytechnic School of Algeria with first class honors (top student). He then joined the University of Minnesota, USA, where he completed his MS and PhD in 1988, and 1992 respectively. He worked for one year as a Post Doctorate Fellow with the University of Minnesota Radiology Department in the area of MRI. He then joined the Queensland University of Technology, Australia, as an Assistant Professor, Associate Professor, then Reader. He joined, in 2001, King Fahd University of Petroleum and Minerals in Saudi Arabia where he is leading the Signal Processing Group. He was awarded the "Best Electrical Engineering Student" award while at the Polytechnique. He has published over two hundred refereed papers in Journals and conferences and has attracted more than $1 Million in research funds. Dr. Deriche supervised 10 PhD students, 5 MS students and been a member of more than 20 PhD and MS committees. He has also supervised more than 60 BS theses. Dr Deriche chaired the IEEE ISSPA conference in 1999. He acted as technical chairman for a number of conferences including: the IEEE TENCON conference in 1997, WOSSPA 1999, IEEE GCCC 2003, IEEE GCC 2004, and the WCBE Signal processing sub-conference in 2003. He delivered a tutorial on wavelets at ISSPA 1999 and at the BTE 2004, and ISSPA 2005. He was a recipient of the IEEE third Millennium Medal in 2000. His research interests are multi-scale signal processing, wavelets, fractals, spectral estimation, with particular emphasis on Multimedia compression applications and biomedical applications.