A New Feature-Based Biometric Identification System in IoT-Powered Smart Cities Using a Hybrid Optimization Algorithm

The pervasive expansion of the Internet of Things (IoT) year after year has facilitated widespread sensing capabilities across various domains. Consequently, heightened concerns have arisen regarding authentication and security measures. There is a growing focus on biometrics in the realm of human identification applications, particularly in the context of advancing biometric-enhanced IoT applications. This trend is garnering increasing attention as it unfolds. As new technologies have developed, biometric-based identification has been seen as an effective way to automatically identify people because of its uniqueness and impossibility of fabricating it. Biometric identity systems make secure authentication and access management possible. However, due to their almost identical physical traits, one of the biggest problems with conventional systems is being able to tell between identical twins. This fact frequently results in high false acceptance rates, putting the system's security at risk. Thus, a solution is addressed in this work by applying a multi-biometric identification method based on unique feature levels. Moreover, the accuracy and robustness of the biometric identifications are further enhanced both with Real Coded Genetic Algorithm (RCGA) and Fish Swarm Algorithm (FSA). RCGA is employed as a global search to explore the promising solution space and guide the solution toward the global optimal region. The algorithm exploits the capability of AFSA, serving as a local search to secure the final optimal solution. Besides, the proposed method enhances the system's discriminative power, enabling an identification more precise and trustworthy. Therefore, this work greatly contributes to the advancement of biometric identification systems and the increase of accuracy in various fields.