Palmprint Recognition Using Deep Learning

Abstract

Biometry is the use of biological features in order to identify individuals; biometric systems aim to make such identification automatic, fast and highly reliable; a biometric system uses one or multiple modalities to verify the identity of a user, a highly recommended biometric modality is the palmprint. Palmprint recognition systems are highly reliable because of the properties of palmprints from their distinctiveness to their stability over time, and also their accessibility since there are multiple ways to acquire a palmprint image. The recognition process goes through multiple steps from the image acquisition, pre- processing, feature extraction and finally matching (or classification); each step has a plethora of ways it can be done by. The recognition of the palmprint is best done using deep learning, using the convolutional neural network AlexNet to extract the features and then classifying the palmprint accordingly, the system overall returns good results and high accuracy.