An Innovative Method for Securing QR Codes against Counterfeits in Supply Chain Management

QR codes have become increasingly integrated with blockchain-based digital supply chain management systems. The immutable and decentralized nature of a blockchain-based system improves transparency among stakeholders. However, there remains the issue of counterfeiting QR codes, as the hash of a counterfeit copied or reprinted QR code is the same as the original. To bridge this security gap, we take advantage of the fact that at a microscopic level, clear differences exist between authentic QR codes and counterfeit copies. The printing process at the microscopic level reveals consistent traces that are non-copyable by counterfeiters. By collecting low-cost microscopic images of QR codes as training data, a Convolutional Neural Network (CNN) can be trained to distinguish between originally printed and counterfeited QR codes. We also propose a dual-channel blockchain system by adding an additional covert channel embedded into the QR code through cell width adjustments. This covert channel may significantly enhance the versatility of the supply chain management system, providing a separate channel for potential security enhancement. In our experiments, the Convolutional Neural Network (CNN) for distinguishing between originally printed and copied QR codes consistently achieved reliable results, with 96% accuracy, 95% precision, and a 98% F1 score after fine-tuning for 50 epochs with AlexNet as the base model. Our simulation results have successfully demonstrated the validity of the proposed method for securing QR codes in supply chain management using CNN in conjunction with a digital microscope and a dual-channel blockchain mechanism.