Image watermarking based on non-Newtonian effect and interpolated SWT-DWT

This paper proposes a robust image watermarking method based on the non-Newtonian effect in fluid dynamics and multiple interpolated transformations. First, the host image is encrypted by using a k-level Arnold map, followed by Discrete Wavelet Transformation. Second, the watermark is inserted into the encrypted-transformed host image, where the resulting image is inversely transformed by using inverse stationary wavelet transformation, followed by the inverse Arnold mapping (decryption) to produce the watermarked image. Specifically, inserting watermark into the Arnold domain diffuses the watermark information throughout the host image, which leads to the non-Newtonian effect. All three color channels of the host image are processed differently to achieve robust watermark extraction and quick tampering detection. Experiments are conducted to verify the basic performance of the proposed method. Results suggest that the proposed method outperforms the conventional work in terms of image quality while achieving high robustness and embedding capacity.