Enhancing Cybersecurity by Deep Learning Models for QR Code Image-Based Attack Detection Using Lion Optimization Algorithm

Abstract

In today's digital landscape, the protection of sensitive data against rising cyber threats is paramount. Traditional methods for detecting cyber-attacks often fall short due to their limitations in accuracy and efficiency. This study proposes a pioneering solution: a hybrid approach integrating lightweight deep learning models like MobileNetV2 and ShuffleNet V2 with the Lion Optimization algorithm for feature selection. By leveraging these advanced technologies, the proposed method addresses the shortcomings of conventional approaches. It significantly enhances both the accuracy and efficiency of cyber-attack detection processes. Notably, the hybrid model achieves an impressive 99.82% accuracy in classifying and detecting various types of cyber-attacks, surpassing the performance of traditional CNN models.