Finger Knuckle Print Recognition using Transfer Learning

Recent years have seen a surge in interest in biometric technologies. The finger-knuckle print (FKP) is one of the most common biometric characteristics. FKP would be a reliable and trustworthy biometric since the dorsal part of the finger is not accessible to exteriors. This article provides an FKP framework for extracting features from FKP images performed using Transfer Learning Convolutional Neural Networks Models. The proposed methodology is validated using the Hong Kong Polytechnic University’s publicly accessible PolyU FKP dataset. Experiment findings demonstrate that the suggested ResNet and SOFTMAX Classifiers attain an F1 Score of 98.9% when used together. According to the results, the suggested biometric identification system is secure, sturdy, and dependable.