Convolutional Neural Algorithm for Palm Vein Recognition using ZFNet Architecture

Palm vein pattern recognition is one of the among biometric recognition techniques that uses blood vessel traits for person's identity identification and/or verification. This paper presents a preliminary study on palm vein recognition based on the application of Convolutional Neural Network (CNN) using a deep learning architecture called ZFNet. ZFNet was adapted and implemented in the proposed method by proposing an improved architecture based on optimal parameters. The proposed method was assessed on the near-infrared palmprint images from MS-PolyU database. The experimental results carried out have shown the high recognition performance of the proposed method compared with other CNN architectures considered in the proposed study such as LeNet, AlexNet and ResNet.