earticle

영어

The QR Code was initially developed to monitor the manufacturing processes of automotive components. However, its application has expanded significantly across various industries over time. QR codes can be created through a range of methods, including online tools, mobile applications, and programming libraries. They are now widely used in several domains such as payment transactions, identity verification, and accessing product information. Despite the convenience offered by QR Codes, they are susceptible to risks, particularly the emergence of counterfeit QR codes. This article presents a novel classification algorithm that leverages the DenseNet architecture to identify fraudulent QR codes. By integrating multiple DenseNet layers (referred to as DenseLayers) atop the standard DenseNet framework, the algorithm enhances the classification model's efficacy. Additionally, we introduce a publicly available QR code dataset that employs the Mish activation function instead of the conventional ReLU activation function. Experimental evaluations reveal that the modified model achieves an average accuracy of 99.8%, representing a 0.5% improvement over the baseline model.