back to Biometrics (CSE 40537/60537)

Progress

_____________________________
Tuesday, 05/02:
Today we have discussed LBP as an example method to extract hand-crafted features for face recognition. Also an alternative methods based on deep learning were discussed. At the end we saw an example application of biometrics in automatic border control. Slides: "Face recognition" (7-end), "Example large-scale application of biometrics: e-passports" (MRZ and BAC).

_____________________________
Thursday, 04/27:
Today we have finished statistical evaluation of biometric systems and started to talk about face recognition. Slides: "Statistical evaluation of biometrics" (26-end), "Face recognition" (1-6).

_____________________________
Tuesday, 04/25:
Today we have finished discussing the security aspects and started to talk about evaluation of biometric systems. Slides: "Security of biometrics" (114-end), "Statistical evaluation of biometrics" (1-25).

_____________________________
Thursday, 04/20:
Practical class No. 5 (acquisition of iris spoofs)

_____________________________
Tuesday, 04/18:
Today we have discussed the security of transmission channels in biometric systems. Slides: "Security of biometrics" (97-113).

_____________________________
Thursday, 04/13:
We have discussed iris liveness detection. Slides: "Security of biometrics" (67-96).

_____________________________
Tuesday, 04/11:
We have discussed fingerprint liveness detection. Slides: "Security of biometrics" (35-66).

_____________________________
Thursday, 04/06:
Guest speakers: Dr. Richa Singh and Dr. Mayank Vatsa on deep learning for face recognition (129 DeBartolo). Right after the lecture: acquisition session in 355C Fitzpatrick ("gummy fingers").

_____________________________
Tuesday, 04/04:
We have discussed cepstral features and finished speaker recognition. And we have started to talk about security of biometrics. Slides: "Speaker recognition" (31-end), "Security of biometrics" (1-34).

_____________________________
Thursday, 03/30:
Continuation of speaker recognition: formants and estimation of features related to our vocal tract (use of auto-regressive models and blind deconvolution). Slides: "Speaker recognition" (9-30).

_____________________________
Tuesday, 03/28:
We have finished discussing how to apply dynamic time warping in signature recognition. And we moved to the next biometrics: speaker recognition. Slides: "Recognition of handwritten signatures" (46-end), "Speaker recognition" (1-8).

_____________________________
Thursday, 03/23:
Practical class No. 3 (acquisition of handwritten signatures)

_____________________________
Tuesday, 03/21:
Today we have discussed the global and local features possible to be used in off-line and on-line signatures. We have also started to discuss how dynamic time warping can be used to compare two signatures. Slides: "Recognition of handwritten signatures" (28-45).

_____________________________
Thursday, 03/09:
Today we have discussed the difference between off-line and on-line signatures and the methods of their acquisition. Slides: "Recognition of handwritten signatures" (1-27).

_____________________________
Tuesday, 03/07:
Today we have finished biometric modes based on hand. Next stop: handwritten signature recognition. Slides: "Use of hand in biometrics" (22-end).

_____________________________
Thursday, 03/02:
Today we have started to discuss the most important modalities related to hand. Also, as an important digression, we used a hand geometry to discuss the biometric system pipeline, transformations between spaces, and classification. Slides: "Use of hand in biometrics" (1-21).

_____________________________
Tuesday, 02/28:
Today we have discussed statistical properties of the iris code and we have finished iris recognition. Next stop: hand biometrics. Slides: "Iris recognition" (78-81 and 83-end).

_____________________________
Thursday, 02/23:
Today we have discussed iris coding: filtering operation and kernels, convolution vs. correlation, quantization of filtering results, calculation of normalized Hamming distance and inclusion of mask into the calculations. We also discussed how to compensate for eyeball rotation. Slides: "Iris recognition" (76, 77, 82, and 83).

_____________________________
Tuesday, 02/21:
Practical class No. 2 (iris data acquisition)

_____________________________
Thursday, 02/16:
Today we have discussed iris segmentation, starting from basic Daugman's idea of using circular approximations of pupil and iris, and then moving to more general (nonlinear) models employing Fourier expansion and active contours. Slides: "Iris recognition" (52->74).

_____________________________
Tuesday, 02/14:
Today we discussed iris image acquisition and representation. Especially, we have learned that an iris image appropriately cropped and masked can be very efficiently compressed and still serve as a good biometric sample. Slides: "Iris recognition" (36->51).

_____________________________
Thursday, 02/09:
Today we have discussed iris genesis and a few hypotheses about its uniqueness and stability. Also, we have started to discuss the iris image acquisition. Slides: "Iris recognition" (1->35).

_____________________________
Tuesday, 02/07:
Today we discussed how to locate the minutiae and how to compare maps of minutiae. And we have finished fingerprints. Slides: "Fingerprint recognition" (72->end). Next stop: iris recognition.

_____________________________
Thursday, 02/02:
Practical class No. 1 (fingerprint data acquisition)

_____________________________
Tuesday, 01/31:
Today we continued discussing fingerprint image acquisition. Also we have discussed shortly the basic principles of wavelet transform used in fingerprint image compression and quality enhancement. We have also discussed various ideas for image segmentation and started to talk about fingerprint matching. Slides: "Fingerprint recognition" (43->71).

_____________________________
Thursday, 01/26:
Today we have discussed fingerprint features: 1-level singular points used in classification, 2-level minutiae used in recognition and tiny 3-level features used mainly in liveness detection. We have discussed Henry's classification and a basic method of singular point detection (Poincare index). Also, we have started to discuss methods of fingerprint acquisition. Slides: "Fingerprint recognition" (10->42).

_____________________________
Tuesday, 01/24:
Today we have discussed tasks of the biometric system (enrollment, authentication), biometric errors (FMR, FNMR, EER) and started to talk about fingerprint recognition. Slides: "What is biometrics?" (31->end), "Fingerprint recognition" (1->9).

_____________________________
Thursday, 01/19:
Today we have discussed biometric modes, the expected properties of biometric characteristics and (very briefly) history of biometrics. Slides: "What is biometrics?" (14->30).

_____________________________
Tueasday, 01/17:
Today we have discussed: course syllabus, definition of biometrics and basic biometric vocabulary (biometric characteristics, sample, feature, template). Slides: "Syllabus" (all), "What is biometrics?" (1->13).