연세대학교 전기전자공학부
페이스북 트위터

세미나 [06/14] Fingerprint Recognition of Young Children 2017.06.02 302

< BK21 플러스 BEST 정보기술 사업단 세미나 개최 안내 >

 

개최일시 : 2017 06 14 수요일 11:00 ~ 12:30

개최장소 : 2공학관 701

세미나 제목 : Fingerprint Recognition of Young Children

발표초록 :

There is a growing requirement worldwide to recognize young children (newborns and infants) for applications such as vaccination tracking, healthcare, delivery of food and medicine, and national registry. In 1899, Galton first captured ink-on-paper fingerprints of a single child from birth until the age of 4.5 years, manually compared the prints, and concluded that the print of a child at the age of 2.5 years would serve to identify him ever after. Since then, new technologies for digital capture of fingerprints and automatic fingerprint comparison have become available and yet only a few studies on child fingerprints have been conducted. We present a detailed longitudinal study that addresses the following questions: (i) Do fingerprints of infants possess the salient features to recognize them? (ii) At what age can an infants fingerprints be captured for recognition? (iii) Does a fingerprint persist over time? We collected fingerprints of 309 children (0-5 years old) at four time instants over a one-year period. We show, for the first time, that (i) fingerprints of six hour old newborn exhibits sufficient information for recognition, and (ii) state-of-the-art matching algorithms can achieve 98.9% TAR @ 0.1% FAR for 6-months and older children. A mixed-effects statistical model shows that the recognition accuracy is not significantly affected for the longitudinal data over one-year time interval. Our study demonstrates the viability of fingerprint recognition for children as young as 6 months.

 

 

강연자 : Anil K. Jain, Professor / Department of Computer Science & Engineering, Michigan State University

초청자 : 전기전자공학과 교수 김재희


목록
prev 세미나 [06/23] New Approaches for III-V Photovoltaic and Photoelectrochemical Solar Energy Conversion 2017.06.22
prev 세미나 [06/08] 5G – Massive, Ultra-Dense, Delay, and Computing 2017.06.02