Deep Learning Technique for Fingerprint Recognition
Abdulmajed Aboalhsan, Mohammed Naif Alatawi · 2022
The use of biometric and special measurements Fingerprints, to verify the identity of users have become effective and reliable, knowing that most of the world's institutions cannot do without biometric software and hardware for ease of use and low operating costs. Biometrics is an advanced part of current society. In this study, a system for fingerprint recognition and identification was designed and implemented. This system was mainly depending on the fingerprint minutiae that represent discontinuous regions in ridges. Deep learning in this study was used to perform the matching process between fingerprints, as it allows the machine to learn itself depending on the large volume of datasets. In this project a method for fingerprint recognition using deep learning is presented. By using CNN, training was made on a group of fingerprints belonging to different people. Eight pictures were passed to each person, in different settings, for easy identification afterwards of any fingerprint. The training process was successful, starting with 75, 98 and 100 percent increase by increasing epoch number. The model was also tested and it succeeded in detecting a fingerprint that was passed on to the model.