A Robust Copyright Protection Scheme Based on Discrete Cosine Transform and Image Normalization

Watermarking is a way that is used to protect the copyright of images. As the existing watermarking methods are not robust enough a new novel scheme is needed. In this paper, we propose a zero-watermark discrete cosine transformation-based copyright protection scheme, in which a normalization procedure is applied in both watermark embedding and extraction to make our scheme more robust to geometric distortions. This method is resistant to malicious signal processing and geometric attacks such as blurring, JPEG compression, noise adding, sharpening, scaling, rotation and cropping. Experimental results on images with different complexity show that our proposed scheme outperforms related methods in most cases. Moreover, our proposed method is lossless, i.e., with no modification on original image, and does not require original image for watermark extraction.