Fingerprint Recognition Introduction - Biometric technology is a way of identifying and verifying individual people by analyzing biological characteristics that are unique to them - Everyone has different fingerprints, even if they are identical twins - So we can use fingerprints as unique biological characteristics -> fingerprints recognition - According to ancient Chinese documents, there is a record of using fingerprints to authenticate documents. - In modern times, fingerprints have been used to select suspects from criminal investigations - 1999s, FBI made Integrated Automated Fingerprint Identification system (IAFIS) - Nowadays, fingerprints recognition technology can be found in many places, ranging from government offices, office access authentication, mobile unlocking, and financial payment simple authentication Morphological Characteristics of Fingerprints Ridge The part of fingerprints that appears in a line and rises like a mountain range Valley The hollow part between the ridges Ending Point The point at which a ridge flows smoothly and ends Bifurcation The point at which a ridge flows smoothly and splits Upper core The largest upward bend of the ridge Lower core The largest downward bend of the ridge Left/Right Delta The point that three flow of fingerprints with different direction converge Unlike general pattern recognition, fingerprint recognition is characterized by a mixture of fingerprint-specific noise Minutiae : non-continuous points that break the constant flow of the fingerprint’s ridges (Galton) Minutiae-Based Fingerprint Identification System (1) Binary : divide object(1) and background(0), and extract direction (2) Smoothing : in previous step, we made sub-intervals to extract some information. In this step, we connect each sub-interval and make it smoothly connected (3) Thinning : extracting only one line from binarized and smoothed line 1. Optical Fingerprint Recognition - Shoot strong light upward where the fingerprint touches, then the fingerprint shape is reflected as it is to optical scanner and images can be extracted from an it. Finally, the image is digitized to determine whether it is the same as the original data. - Good durability, Cheap to make / recognition rate is relatively low 2. Capacitive Fingerprint Recognition - When you put your fingerprints on this sensor, there will be distance difference, because of ridges and valleys. Because of difference in capacitance due to difference in distance, the fingerprints can be recognized. - There are two types of method (area method vs swipe method) - Area method is easier and faster than swipe method, but it is difficult to miniaturize sensnors and the unit price is high 3. ultrasonic fingerprint recognition - This method measures the return time by emitting ultrasonic waves, and recognizes the fingerprint by measuring the height difference of the fingerprint. - More secure than optical fingerprint recognition, because it recognizes the user’s fingerprint curvature, not the fingerprint image. - It also can penetrate objects. Using this property, fingerprint recognition has become possible even under the display. - Good accuracy and durability / difficult to manufacture and relatively expensive Reference Samsung display newsroom https://news.samsungdisplay.com/18221 “생체측정을 이용한 안전한 인증 프로토콜 설계 및 효율적 지문인식 알고리즘의 구현” https://postech.dcollection.net/public_resource/pdf/000001905712_20230506210657.pdf
