Enhancing fingerprint identification using Fuzzy-ANN minutiae matching

‘Based on Minutiae and Neural Networks,’ this paper introduces a robust fingerprint identification system that significantly enhances the accuracy of matching fingerprints, especially those altered due to various reasons such as scars or mutilations. Utilizing a combination of minutiae-based matching and neural network algorithms, the system is designed to overcome the limitations of traditional methods, which often fail under less-than-ideal conditions. The system's core lies in its ability to train an artificial neural network to learn an improved similarity function for minutiae matching. This capability has been extensively validated through a series of rigorous experiments, demonstrating its superiority over existing systems. Implemented in MATLAB, the system performs remarkably on benchmark datasets like FVC2004 DB1 and NIST SD27, achieving state-of-the-art results. This paper not only presents a detailed methodology involving image enhancement, minutiae extraction, and advanced matching techniques but also sets a new standard in fingerprint identification technology, particularly in handling altered fingerprints effectively.