Biometrics & Security Tutorial 7 • 1 (a) Please compare two different kinds of biometrics technologies: Retina and Iris. (P8:2-3) • 1 (b) Understand the retina recognition system. Please give the differences between retinal identification and retinal authentication. (P8: 10-11) • 1 (c)P8:20 show a representative example of identification system. Please give another example of verification system. • 1 (d) Please understand the iris recognition system. (P8: 22-36) • 1 (e) What is texture feature (P8: 30)? Why we can use texture feature in iris recognition (P8: 31)? • 2. There are four steps in the Daugman’s approach (P8: 32-36). The third step generates IrisCode with 512 bytes. If 2 bits represent a feature, please compute the total number of features. (512*(8/2)=2,048) • Hamming distance: the distance between two vectors A and B is ∑ | Ai - Bi |. General idea of Iris Code • • • • • 3. In P8:37, an example of iris verification distributions is given. Notice that Hamming distance is defined to measure the similarity of two IrisCodes. Which value in the figure can separate two parts, authentics and imposters? If the similarity between two IrisCodes is taken as 0.5, can you say they are identical? How about the comparison result when the similarity is 0.2? What conclusion will you get from the hamming distance used to compute similarity of IrisCodes? (0.38; no; yes) True Acceptance Rate • A genuine individual is accepted. (TAR) TAR Total True Acceptance Total True Attempts Threshold=0 means that the attempts whose matching score >0 will be accepted; (everybody will be accepted) 1 TAR 0 Threshold (Matching scores) Threshold=∞ means that the attempts whose matching score > ∞ will be accepted (nobody will be accepted) False Rejection Rate • A genuine individual is rejected. (FRR) Total False Rejection FRR Total True Attempts Threshold=0 means that the attempts whose matching score <0 will be rejected; (nobody will be rejected) 1 FRR 0 Threshold (Matching scores) Threshold=∞ means that the attempts whose matching score < ∞ will be rejected (everybody will be rejected) True Rejection Rate • A impostor is rejected. (TRR) TRR Total True Rejection Total False Attempts Threshold=0 means that the attempts whose matching score <0 will be rejected; (nobody will be rejected) 1 TRR 0 Threshold (Matching scores) Threshold=∞ means that the attempts whose matching score < ∞ will be rejected (everybody will be rejected) False Acceptance Rate • A impostor is accepted. (FAR) FAR Total False Acceptance Total False Attempts Threshold=0 means that the attempts whose matching score >0 will be accepted; (everybody will be accepted) 1 FAR 0 Threshold=∞ means that the attempts whose matching score > ∞ will be accepted (nobody will be accepted) Threshold (Matching scores) ROC • There are some trade off among TAR, FRR, TRR, FAR. • We need a particular threshold to keep the TAR and TRR as high as possible and to keep FAR and FRR as low as possible. • We need the ROC curve to find the optimal threshold. • We also use ROC curve to evaluate the system ROC(2) Find the optimal point (threshold) ROC Curve ROC Curve 1 1 TAR FAR 0 FAR 1 0 FRR 1 • 4. P8:27 shows the corresponding Fourier transforms of iris images with different quality. Please give the property of Fourier transform of good quality iris images. (Even distribution) Fourier Transform • Any signal can be expressed as a weighted sum of a series of sine or cosine wave • a is the original signal • a = w1∙b + w2∙b + w3∙d • b is a low frequency component • d is a high frequency component A sum of sines and cosines = 3 sin(x) A + 1 sin(3x) B + 0.8 sin(5x) C + 0.4 sin(7x) D A+B A+B+C A+B+C+D • 5. In Daugman’s approach iris is regarded as a circular pattern (P8: 3236). In fact we can also take the iris as other pattern, as shown in P8: 29. Please try to design the corresponding scheme.