| It is known that the fidelity of the image is inversely proportional to the robustness of the watermark. Therefore, there should be the trade off between fidelity for robustness and vise versa. Based on the new spatial masking we explored in this thesis, a new adaptive digital image watermarking method is proposed. It keeps the invisibility of the watermark and maintains its robustness at the same time. The new spatial masking is built according to the image features such as flat areas with big brightness or darkness, edges, and regions with high activities. With the same watermarking energy, the quality of water-marked image with this masking is much better than the one without it. We also propose using the weighted Peak Signal-to-Noise Ratio (wPSNR) to evaluate the image quality. The watermark is detected by the key-dependent method without knowing the original image information. In addition, we extend this proposed spatial masking to the Discrete Cosine Transform (DCT) domain by using the method of searching the extreme value of the quadratic function subject to the bounds on the variables. |
