Print-Cam Robust Image Watermarking Based on Hybrid Domain

Image watermarking is an effective means of hiding and reading information. However, the process of printing out an image and then taking a picture of it, also known as the print-cam process, introduces major noises such as geometric distortion and picture blurring as well as color degradation. It is a strong attack party. Digital image watermarking is often not robust to print-cam attacks. In this paper, we propose an effective image watermarking method against print-cam attacks embedded in the composite domain of the image. The B channel of the image is selected for watermark embedding. Firstly, the Discrete Wavelet Transform (DWT) transform is performed on the image, and then the blocks in the LL sub-band are taken and the Discrete Cosine Transform (DCT) is performed, finally the embedding of watermark information is completed by modifying the middle frequency in the frequency domain. When extracting the watermark, Scale-Invariant Feature Transform (SIFT) feature points are firstly used to calculate the image feature points, and then the matching calculation between the feature points is calculated to get the perspective transformation matrix of the photographed image and corrected. The experimental results show the method has the ability to resist the print shooting attack when the watermark has less effect on the image quality. The means of adjustment is also proposed to ensure the wide applicability of the algorithm, which can meet the needs of many application scenarios in real life.