Robust Detection of Copy-Move Forgery in Color Images

With rapid advances in digital image processing technology, there is a massive development of sophisticated tools and techniques for digital image forgery. Copy-move forgery is one of the techniques most commonly used. A part of an image is copied and pasted into the same image, in order to maliciously hide or clone an object or a region. In this paper, we propose a method that addresses this kind of fakery. Each component of the suspicious RGB image is used for feature extraction. Let Y be the Y-component of the corresponding YUV image. Then R, G, B and Y are splitted into fixed-size overlapping blocks, and characteristics are computed from each component. The 10-dimensional feature vectors made by concatenating the characteristic values of each component are then stored row wise in a matrix that will be lexicographically sorted to make similar image blocks neighbors. Duplicated image blocks are identified using Euclidean distance as similarity criterion. Experimental results showed that the proposed method can detect the duplicated regions even in case of slight JPEG compression, blur and noise addition.