Blur Detection Of Tampered Image

As one of the most important message carrier, digital image has been widely usedin the age of information and the Internet. And with the development of informationtechnology, a variety of powerful and practical digital image editing software continueto be used. But many malicious tampered images which are made by people havebrought a negative influence to the society. So the information security of digitalimage has become a big problem which needs to be resolved.This article through in-depth analysis of blur tampering and the principle of blurtampering. In this article, we have decided to put the blur tampering into twocategories, namely local blur tampering and overall blur tampering. And we havemade a simple introduction of some existing detection algorithm of blur tampering.The focuses of this article are detection of local blur tampering and detection ofoverall blur tampering.For the detection of local blur tampering, this article first analyzed a methodwhich is based on the homomorphic filtering. And in the further study, we used thewavelet homomorphic filtering and a kind of homomorphic filtering that has beenimproved to replace the traditional homomorphic filtering in the detection, and wehave compared the effect of three types of filtering. This method can effectively foundthe tampered regions, and we found the detection accuracy is improved by themodification of the homomorphic filtering.For the detection of overall blur tampering, this article first analyzed the principleof blur tampering, and we have proved that the correlation between image pixel andits neighborhood pixels will be enhanced after the image was tampered. On the basisof this theory, we proposed a method can use to detect blur tampering which is basedon the correlation between image pixel and its neighborhood pixels. This method can effectively and quickly judged an image whether it is overall blur tampering.