Smart Phone Based Online Medicine Authentication Using Print-Cam Robust Watermarking

摘要

In this paper, we have proposed a print-cam resilient watermarking application for online authentication of a medicine using a smartphone. Print-cam attack induces geometric distortions during the image capturing process. This watermarking application is able to blindly retrieve the information from the image captured using a smartphone camera. Watermark message is first coded as periodic directed pattern, and then embedding is carried out using additive method. At the extraction side, we have introduced corner detection to revert the geometric distortion caused by print-cam process and then the regularities in autocorrelation function of the periodic pattern are utilized. The robustness against distortions from printing and camera lens is ensured using noise reduction, modified JND model, and two-level coding of the message. Experimental results show that the proposed algorithm is robust against print-cam attack and can be used for potential camera based online medicine authentication.