PROJECT TITLE :

Domain Specific Learning for Newborn Face Recognition

ABSTRACT:

Biometric recognition of newborn babies is an opportunity for the conclusion of many helpful applications, such as improved security against swapping and abduction, correct census, and effective drug delivery. This paper explores the possibility of using face recognition toward an inexpensive and friendly biometric modality for newborns. The paper proposes an autoencoder-based mostly feature representation followed by downside specific distance metric learning via one-shot similarity with one category-online support vector machine. The largest publicly out there database of newborns collected from varied sources to check face recognition is introduced. Many existing face recognition approaches and business systems are also evaluated on a common benchmark protocol. The efficacy of the proposed algorithm is evaluated beneath both verification and identification settings. With multiple galleries, rank-1 identification accuracy of 78.five% and verification accuracy of 63.four% at 0.1% false settle for rate are achieved.