OLFV: Harnessing the Power of Enhanced Deep Learning Model to Recognize Fingerprints using Optimization and Classification Principles Paper

Among many forms of biometric identification, fingerprints are among the most common. Due to their individuality, they are thought to be vital in forensics. Human operators have long been responsible for person identification. On the other hand, automatic fingerprint identification systems have emerged thanks to technology advancements, and their robustness is becoming more important as the population grows. The field of computer vision, however, has benefited greatly from deep learnings advancements. Optimized Learning for Fingerprint Verification (OLFV) is a new method for accurate fingerprint identification that is introduced in this paper. To test its efficacy, it is cross-validated with a traditional deep learning model, Convolutional Neural Network (CNN). While fingerprint analysis falls under the purview of computer vision and image processing, deep learning has seen relatively little use in this field thus far. Two steps in the automated fingerprint identification pipeline-fingerprint categorization and fingerprint minutiae extraction-are the subject of this study use of deep learning techniques. For those two phases, we build and evaluate deep learning systems according to several criteria, including dataset size and network design. A system that uses proposed convolutional models for decision making uses the preprocessed fingerprint photos. With the Optimized Learning for Fingerprint Verification variation reaching over 98.54% accuracy and the CNN variant reaching over 95.26%, the system is robust. The resulting section shows the proper proof to it in clear manner with graphical illustrations.