Semi-blind image restoration based on telltale watermarking

We propose a novel concept for the restoration of locally degraded images based on telltale fragile watermarking. In our approach a data stream called a watermark is embedded in the wavelet domain of an image such that manipulation of the image is reflected in the embedded stream. The altered stream is used for semi-blind restoration to undo tampering. It is assumed that the embedded watermark is known, and the degradation is in the form of localized filtering in which the explicit filter characteristics are unknown. Simulation results demonstrate the potential of the approach for practical tamper recovery.