Research on Pseudo-Random Noise Information Identification Technology of Printed Anti-Counterfeiting Image Based on Deep Learning
Yumeng Zhen, Peng Fei Cao, Feng Liuping, Chen Fangfang · 2020 5th International Conference on Computer and Communication Systems (ICCCS) · 2020
Halftone image with hiding pseudo-random information, which has excellent anti-duplication performance. It's used to against the problem of counterfeiting and copying QR Code. However, due to the mocro-size of pseudo-random image information, direct recognition with mobile phones often has high bit error rate, low reliability and badly user experience. This paper based on the method of deep learning, to through marking the noise data in the QR Code images and classify them. Then building a convolutional neural network model, training the features of the noise image. Proved by experiment, this method can effectively identify noise information and the result can satisfy the error-tolerance requirement of pseudo random noise image channel coding.