Using Crossover Area of the Retinal Eye to Generate a Secret Key Based on Deep Learning- YOLO-v5 Object Detection Algorithm
Zainab Sahib Dhahir, Israa Ali Alshabeeb, Khamael Raqim Raheem, Wafaa Mohammed Ridha Shakir · International Journal of Safety and Security Engineering · 2024
Traditional security keys based on passwords are vulnerable to attacks.Recent developments in biometric security systems offer a solution.This study presents a novel approach for generating security keys using the YOLO-v5 deep learning model to detect crossover points in retinal images.We extract RGB hash values from these points and use blockchain to create secure keys.This method enhances security and reduces processing delays.Our approach achieved an F1-score of 69% to 76% on the DRIVE, IOSTAR, and Ibn Al-Haytham datasets, demonstrating promising results compared to existing biometric security technologies.