Cancelable biometric authentication system based on hyperchaotic technique and fibonacci Q-Matrix

Due to the rapid growth in the field of information transmission techniques, billions of data are transmitted through variant authentication applications between clients. Providing protection to store and transmit these details has the priority in most services. Cancelable biometric techniques have the ability to store and employ these biometric signatures safe away any intruders. These template protection techniques are derived into transformation and cipher based cancelable biometric methodologies. The cipher approaches that are dependent on chaotic schemes quite achieve the two requirements of successful cancelable systems known by confusion and diffusion. This is prompted us to propose a novel cancelable research combines between Hyperchaotic technique and Fibonacci Q-matrix. The hyperchaotic system employs six initial conditions, equations, to construct the initial deformed patterns. Afterwards, the stage of diffusion is accomplished by employing Fibonacci Q-matrix for the resultant dislocated patterns to obtain the final protected template. The proposed cancelable authentication technique is validated by the two evaluation perspectives of authentication and security measurements, respectively. Furthermore, the proposed cancelable system is compared with the most recent studies in the same state-of-art which are utilized by GA and RNA-GA. Five tested biometrics are exploited to prove the validation of our work through any gray or colored biometric patterns with different sizes under different capturing environments even in the presence of noise level. The proposed work achieves high authentication performance by large values of AROC (Area under Receiver Operating Characteristic) Curve leads to 0.999, and similarity coefficients of licensed clients. As well as, the least values of FAR, and FRR (False Acceptance, and False Rejection Ratios) in addition the more uniform resultant histogram shapes due to the cancelable patterns. From the security perspective, the proposed cancelable methodology achieves high entropy of 7.99, NPCR (Number of Pixels Change Rate), and UACI (Unified Average Changing Intensity) values lead to almost optimum results of 99.0271%, and 31.2058%, respectively. This validates its immunity against probable differential attacks.