A Non-intrusive Method for Copy-Move Forgery Detection

The issue of verifying the authenticity and integrity of digital images is becoming increasingly important. Copy-move forgery is one type of image tempering that is commonly used for manipulating digital content; in this case, some part of an image is copied and pasted on another region of the image. Using a non-intrusive approach to solve this problem is becoming attractive because it does not need any embedded information, but it is still far from being satisfactory. In this paper, an efficient non-intrusive method for copy-move forgery detection is presented. The method is based on image segmentation and a new denoising algorithm. First, the image is segmented using a multi-scale segmentation algorithm. Then, using the noise pattern of each segment, a separate noise image is created. The noise images are used to estimate the overall noise of the image which is further used to re-estimate the noise pattern of different segments. The image segments with similar noise histograms are detected as tampered. A comparison with a state-of-the art non-intrusive algorithm shows that the proposed method performs better.