Robust image authentication using hash function

We propose an image monitoring scheme for detecting malicious alternation and broadcast monitoring. This method has a property that a desired absolute error is admitted for a monitor image. In other words, the monitor image is recognized as the suitable image within the desired error. The property allows us to monitor the image robustly against lossy codings, transmission errors, and so on. In this method, we generate a reference image from the monitor image, and then calculate a hash value as feature information. Selecting the generation method of the reference image and the desired error can control the monitoring accuracy and the robustness.