Implementasi Fingerprint Recognition pada Keamanan Folder Menggunakan Algoritma Convolutional Neural Network

Abstract

Data, encompassing text, images, sounds, and symbols, pervades the digital realm, split into public and private spheres. Public data poses no harm when known, while illegal access to private data can result in severe consequences. Traditional data security methods, like passwords, prove less effective with advancing computing capabilities. This study proposes leveraging fingerprint biometrics, specifically utilizing the Convolutional Neural Network algorithm, for enhanced data and folder security. Using a dataset of 740 fingerprints, sourced from 74 participants aged 12 to 71 years, undergoes processing and augmentation, yielding 79,920 fingerprint images. Employing the Siamese model, the research achieves 90% accuracy in distinguishing diverse fingerprint images. To bolster data integrity, the Advanced Encryption Standard algorithm encrypts and decrypts the secured data. The outcome is a Windows-based desktop application, providing a robust solution in the evolving landscape of data protection, encapsulated within this succinct research.