New Performance Evaluation Method for Data Embedding Techniques for Printed Images Using Mobile Devices Based on a GAN

摘要

Methods that embed data into printed images and retrievedata from printed images captured using the camera of a mobile devicehave been proposed. Evaluating these methods requires printing and capturingactual embedded images, which is burdensome. In this paper, wepropose a method for reducing the workload for evaluating the performanceof data embedding algorithms by simulating the degradation caused byprinting and capturing images using generative adversarial networks. Theproposed method can represent various captured conditions. Experimentalresults demonstrate that the proposed method achieves the same accuracyas detecting embedded data under actual conditions.