New robust method for image copyright protection using histogram features and Sine Cosine Algorithm

摘要

Zero-watermarking methods are widely used for efficient copyright protection of digital images. These methods have the ability to withstand both common image processing attacks and some geometric attacks. However, they cannot effectively resist the complex image attacks such as translation, cropping, combined geometric attacks, UnZign, etc. For this purpose, we propose in this work a novel robust zero-watermarking method that can effectively resist several complex image attacks. The proposed method involves the use of the histogram descriptor to generate a secret sequence of binary values from a user-selected Region of Interest (ROI). In order to check the intellectual property rights, the Sine Cosine Algorithm (SCA) is used for detecting the ROI in the attacked image. Then, the computed histogram of this ROI is binarized with a secret method. Next, the resulting sequence of binary values is compared to the original one generated from the original ROI. If there is a high similarity between these binary sequences, the grayscale image copyrights are validated. The simulation results show that the proposed method is not only resistant to geometric attacks (rotation, scaling) and to common image attacks (JPEG compression, filtering, etc.), but it is also robust against the most complex image attacks (cropping, translation, combined geometric attacks, etc.). Furthermore, the results of the comparisons carried out in terms of robustness against different types of attacks prove the superiority of our method over other similar recent zero-watermarking methods.