The College of Education for Pure Sciences in the Department of Computer Science researched a master's thesis on
(Fingerprint Recognition System Using Deep Learning)
The thesis presented by the researcher (Hussein Ghaleb)
Biometric systems play a vital role in modern security and identity recognition systems, as many countries have adopted their use in different fields. These systems rely on the unique characteristics that distinguish each individual, such as facial recognition, retinal scanning, hand geometry, and especially fingerprints. Fingerprints are one of the most common and widely used biometric methods throughout history, due to their ease of use and high reliability. They also have unique properties and as a result, fingerprints have found wide applications in many fields such as access control, law enforcement and other fields. With the increasing use of these technologies in daily life,
The aim of the thesis
is to provide solutions to these challenges by integrating deep learning techniques and employing different models to improve system performance. Despite the changes that may occur in the quality of fingerprints due to multiple factors such as skin condition, degree of pressure and missing parts of the fingerprint. In this paper, different deep learning models, including visual engineering ensemble, deep convolutional neural networks known as , and the proposed convolutional neural network were integrated with the aim of improving the accuracy of the fingerprint recognition system. The performance of the models was evaluated using a dataset where the models achieved excellent results with a score of However, to further improve the reliability and accuracy of the system, a fusion technique based on predictions from the trained models was applied, which resulted in a significant improvement in the performance with the combined system reaching a score of . In addition, the system was implemented on an architecture to make the proposed solution practical and suitable for use in real-world applications.
