Hybrid Digital Watermarking Scheme Based on Discrete Wavelet Transform for Copyright Protection

TLDR: The proposed method may become robust against the different attacks, such as JPEG compression, salt and pepper noise, Gaussian noise, speckle noise, Poisson noise, contrast adjustment, and different filtering, etc.

Abstract: The quality of the watermarking technique greatly depends on the selection of the watermark structure and insertion strategy. In the problem of watermarking, the involved two main constraints are those of maintaining the robustness of the watermark information while keeping the visual perception of the original image intact. For this purpose, the most important properties are robustness and invisibility. The main objective of the proposed algorithm is to improve the perceptual quality of the watermarked image without violating its robustness against the different attacks. The host image is being divided into four non-overlapping multi-resolution sub-bands by applying Discrete Wavelet Transform (DWT). In the proposed algorithm, horizontal detail region of the host image is being chosen for embedding the watermark. The DWT is also applied to the watermark and its approximation region is chosen to embed into the host image as this region has the most visual perceptual quality for human eyes. At the same time, the approximation of the host image found after 3-level DWT is added to the approximation region of the watermark before embedding. As the level of DWT increased, an image is being decomposed into much lower frequency. The approximation after 3-level DWT contains a very low frequency and high amplitude value. So, in this region image, quality is almost same to the original image. This will make the host image more noticeable than the watermark image and increase imperceptibility of the watermark. A secret key is used for embedding and the same key is used in the extraction process. The proposed method may become robust against the different attacks, such as JPEG compression, salt and pepper noise, Gaussian noise, speckle noise, Poisson noise, contrast adjustment, and different filtering, etc.