Pixel-based hybrid copy move image forgery detection using Zernike moments and auto colour correlogram

摘要

All of us rely on the images for the memories of our life and loved ones. The images are useful in proving an evidence of the event. Image forgery has become much prominent nowadays and is being done either for fun or intentionally. In this paper, we detect copy move forgery by combining the two features namely, Zernike moments and auto colour correlogram. The first checks the shape of the objects in the block, the latter for distance of each colour pixel taking into account the 64 colours. Both combine together to identify the regions for which copy-move forgery exists. The first completes the number of comparisons in the order of n~(2), where n is the number of blocks. The ACC verifies the colours used in the block. If both matches are confirmed, the suspicious blocks are returned. The method out-performs the existing methods based on the probability approach.