Proposed Biometric Security System Based on Deep Learning and Chaos燗lgorithms

Nowadays, there is tremendous growth in biometric authentication and cybersecurity applications. Thus, the efficient way of storing securing personal patterns mandatory most governmental private sectors. Therefore, designing implementing robust security algorithms for users’ biometrics still a hot research area to be investigated. This work presents powerful system (BSS) protect different modalities such as faces, iris, fingerprints. The proposed BSS model based on hybridizing auto-encoder (AE) network chaos-based ciphering algorithm cipher details stored ensures their secrecy. employed AE unsupervised deep learning (DL) structure used extract main features. These obtained features are utilized generate two random chaos matrices. first matrix permute pixels images. In contrast, second further confuse resulting permuted using two-dimensional (2D) chaotic logistic map (CLM) algorithm. To assess efficiency BSS, (1) standardized color grayscale images examined fingerprint, iris were (2) comprehensive recognition evaluation metrics measured. assessment results have proven robustness superiority compared other existing models. For example, succeeds getting high under receiver operating characteristic (AROC) value that reached 99.97% low rates 0.00137, 0.00148, 0.00157 equal error rate (EER), false reject (FRR), accept (FAR), respectively.