||In recent years, as network being more and more developed, the digital multimedia, such as image, sound, film, etc. are conveyed in the internet network, and are easier to be duplicated and usurped illegally. Hence, the authentication of the property ownership of intelligence and the verification of the image integrity seem extremely important in the network environment. The most frequently used, effective technology to solve this problem is the digital watermarking. By digital watermarking, it not only can verify the integrity and ownership of the digit image but also can find out the position and size that are altered.
In this thesis, we propose a scheme of automatic image authentication and recovery. We use an invisible fragile watermarking to verify the image. Before doing the digital watermarking, we choose the region of important (ROI) in the original image and perform fractal image encoding for range blocks of ROI. And then we insert the fractal codes and the watermark into the original image and obtain a watermarked image. By watermark extraction, we can determine whether the watermarked image has been altered or not. If the watermarked image has been altered, we can automatically find out the position and size of the altered region, and apply fractal image decoding and image inpainting to precisely recover them.