Biometrics

(3 Credit Hours) · Undergraduate

The aim of this course is to introduce the principles of automatic biometric authentication. The course will study those biometric modalities which have commercial implementations (such as fingerprints, face, iris, voice, finger veins, handwritten signatures), as well as emerging techniques (such as brain or thermal signals). We will discuss hopes, fears, limitations and strengths related to the presented modalities, including biometric data aging, "reverse engineering" of biometric templates or possibility to use biometrics in post-mortem forensic analysis. Important part of this course will be security of biometrics (in particular presentation attack detection) and secure biometric implementations. Where appropriate, current large-scale deployments of biometrics, such as NEXUS program or biometric passports, will be used as illustration of problems discussed in class. The course will also show how to apply statistics for biometric reliability evaluation in a mathematically elegant way. During five practical classes students will interface with up-to-date commercial biometric sensors and collect an authentic biometric data as well as spoofing samples used during homework. For instance, using various computer vision tools and software libraries students will build their own iris, fingerprint and signature recognition systems following real-world implementations of these modalities, additionally resistant to various attacks such as presenting irises printed on a paper or gummy fingers. Practical classes will utilize the software designed in MATLAB and C/C++.