An efficient and robust method for detecting region duplication forgery based on non-parametric local transforms

Region duplication is one of the most popular methods for digital image forgery, which means that a part of the image is copied and pasted on another part of the same image to conceal undesirable contents. Block matching, mainly based on finding exact or similar copies of pixels, is the common method to detect region duplication. It often suffers time-consuming and is prone to make false decision when the copied region is retouched. In this paper, we describe a new approach with the notion of nonparametric local transform taken as the feature of image block. The experimental results show that the proposed method can reduce the time complexity of detecting algorithm, and is more robust than the traditional methods.