Research On Recognition Method Of Periocular Biometrics

With the continuous development of science and Internet technology application,biometric identification technology has been widely applied in information security and identity verification.Under less constrained environment has become one of the key research areas in biometric recognition in recent years.In this condition,due to the complement of iris recognition and face recognition,periocular recognition has received more and more attention.This paper mainly proposes a new periocular recognition method based on deep learning to improve the performance of periocular recognition and verifies the influence of eyebrow region in the key area of periocular on the performance of periocular recognition.The main research work is as follows:Firstly,In order to improve the performance of periocular recognition,a new method based on Deep Convolutional Neural Networks referred to as Periocular Net was proposed.Periocular Net exploits a 16-layer convolutional neural network,integrated with a residual learning module,and adopts the Arc Face loss function.Data augmentation is introduced in training process.The proposed 16-layer convolutional network framework with residual-learning structure can extract more abundant periocular feature information,Arc Face loss function has better classification ability,Data augmentation can avoid over-fitting caused by insufficient data in the training process.The experiments on two open periocular datasets,which improves the periocular recognition performance compared to the related methods.Secondly,in order to verify the effect of the eyebrow region feature on the performance of periocular recognition in the end-to-end and base on Deep Convolutional Neural Networks approach.two periocular datasets,UBIPr-1 and UBIRIS-1 are established.The periocular datasets includes three cases where the eyebrow area remains unchanged,the eyebrows are removed,and the eyebrow area is filled with skin color.Experimental results show that eyebrow region remained unchanged have the best periocular recognition performance,The second is the eyebrow area filled with skin color periocular data,The periocular recognition performance is the worst when the eyebrow area is removed.The results show that both the color and texture feature of eyebrow region can affect periocular recognition performance and the texture feature information has the greatest influence on the periocular recognition performance.In the endto-end periocular recognition method based on deep learning,which indicates the eyebrow region can improve periocular recognition performance.