Biometric recognition is becoming increasingly important for security applications. This short course is aimed at introducing several algorithms useful for biometric recognition. In particular, common recognition algorithms for face, fingerprint, iris and palmprint biometrics will be discussed. This tutorial will have the following components: motivation for biometric recognition; pattern recognition basics; performance metrics for biometric recognition; biometric recognition algorithms (Principal Component Analysis, Linear Discriminant Analysis, Correlation Filters, Artificial Neural Networks, minutiae-based fingerprint recognition methods, Iris code method for iris recognition and feature-based approaches for palmprint recognition); multi-modal biometric recognition.

As recording densities in optical data storage systems increase, read channels must cope with increased intersymbol interference (ISI) and decreased signal-to-noise ratio (SNR). ISI can be completely eliminated, but this leads to significant reduction in SNR. A more practical alternative is to employ partial response (PR) equalization that aims at controlling the ISI instead of eliminating it. Since PR methods lead to known ISI patterns, optimal data detection can be achieved using maximum likelihood (ML) techniques such as Viterbi Algorithm. In this course, we discuss the basics of PRML and illustrate how PRML is applied to optical data storage.