Detection of Fingerprint Authenticity Based on Deep Learning Using Image Pixel Value

Abstract

Research on fingerprints has been done a lot, this is because of so many uses of fingerprints as an access tool to enter a system. This method is used to ensure the authenticity of authorized users. Fingerprints are used as biometric identification because fingerprints have a unique pattern that is different from every human fingerprint. The many uses of fingerprint biometric systems also cause many threats to the system, fingerprint forgery occurs so that it can be used to access the system illegally. Therefore this study proposes a system to be able to recognize the authenticity of a fingerprint. CNN is generally designed for object recognition of an image, making it suitable for recognizing fingerprint images to determine if a fingerprint is genuine or fake. The results of the evaluation of several experiments conducted obtained the highest accuracy value of 95.32% for determining the authenticity of fingerprints.