A Proposed Cancelable Biometrical Recognition System (CBRS) Based on Developed Hénon Chaotic-Map

Abstract

Nowadays, human biometrics are widely used in authentication systems. In reaction to violent attacks, cancelable biometric patterns are developed from the original templates to increase the security level of biometric characteristics. This study proposes a solution for a cancelable biometrical recognition system (CBRS) based on the created Hénon chaotic-map idea, which increases key space and hence privacy. The suggested CBRS system ensures that the original biometric traits are updated and encrypted before they are saved in the database, protecting them from unwanted cyber-attacks. It makes efficient encryption of face biometric templates possible. The extraction of biometric characteristics is the first step in this design. Following that, the obtained biometric characteristics are encrypted using the suggested model, which causes pixel confusion and diffusion by developing a Henon chaotic map with variable block sizes at different modes of operation. Various face biometrics datasets were used to test the proposed approach. Various metrics, including security and statistical analyses, demonstrate the effectiveness of the approach, including histogram analysis, correlation coefficient analysis, maximum deviation factor analysis, irregular deviation factor analysis, number of pixels change rate analysis, unified average changing intensity analysis, time analysis, and key space analysis. Furthermore, the performance of the proposed approach was assessed using the receiver operating characteristic curve, which was constructed to assess the system's performance. Results of the analysis show that the suggested technique is very effective, resilient, and dependable, as evidenced by its great performance across diverse recognition databases when compared to traditional and modern algorithms, hence improving the security and reliability of biometric-based access management. The proposed method yields an average AROC of around 1, a correlation coefficient of about 0.00013, and an entropy close to one.