A New Type of Two-dimensional Anti-counterfeit Code for Document Authentication Using Neural Networks

A New Type of Two-dimensional Anti-counterfeit Code for Document Authentication Using Neural Networks

연구 분야: Cryptography

학회: ICCSP 2020: Proceedings of the 2020 4th International Conference on Cryptography, Security and Privacy

With the development of technology, the precision of scanning and printing is on the increase. Nowadays, People can easily forge prints with the facility. We need an effective anti-counterfeiting method. Graphical codes are an inexpensive method in the production of anti-counterfeiting. It is generally composed of random-like binary images and printed with high precision. After scanning and reprinting, the information on the pattern will irreversibly be damaged. Thus we can verify the authenticity of the print based on the loss rate of the information of the graphical codes. In this paper, we present a new anti-counterfeiting code, called Two-Level Dot Code (2LD-code), by replacing each dot in the 2D dot code with a particular anti-counterfeiting pattern. This 2LD-code realizes the anti-counterfeiting function while storing extra information. A deep neural network is applied to verify the authorization by extracting the anti-counterfeiting pattern and distinguish original prints and duplicate ones. Experimental results demonstrated the efficiency of the proposed method.