Comparison of Deep Learning and Machine Learning Methods for Fingerprint Verification and Forgery Detection

Today, digital security and authentication have gained critical importance with the rapid development of information and communication technologies. In particular, biometric authentication methods are widely accepted because they are secure and convenient. Fingerprint recognition systems are one of the most common and effective among these biometric methods. However, with the development of technology, security threats such as fake fingerprint attacks also emerge. In this study, an analysis of fingerprint verification and forgery detection using deep learning methods and machine learning techniques is presented. In this project, which was carried out using the SOCOfing dataset, various deep learning models and machine learning algorithms were applied and the accuracy rates of these methods were compared. The main purpose of the study is to improve the fingerprint verification and forgery detection performance with different model configurations and hyperparameter settings. The obtained results were compared using deep learning and machine learning methods and the aim was to determine the most effective techniques.