Ear recognition for newborns

Abstract

Newborn swapping, missing, mixing, and illegal adoption is a global challenge and to resolve this emerging issue very less research has been done. Most of the biometric systems are developed for adults and extremely few of them address the difficulty of newborn recognition. As they are the highly non cooperative users of biometrics the ear of newborn may be a perfect source of data for passive identification of newborn. This paper investigates a unique approach for the automatic recognition of newborn using 2D ear imaging. Our investigation develops a computationally effective and attractive solution to recognize newborn automatically. The proposed algorithm yields Rank-1 identification accuracy of 89.28% on a database of 210 subjects. © 2015 IEEE.