Recaptured Image Detection Based on Texture Features

With the development of digital image processing technology, image capture and image tampering are easy to obtain with the help of portable devices and software tools. Subsequently, digital image forensics has become increasingly important, in which recaptured image detection is one branch. In this paper, a set of features based on image texture are used to identify the recaptured images. Because the recapture process generally accompanies with some image quality losses, which can be reflected from the texture features, we study the effectiveness of LBPV and the proposed Relative-Contrast. Then, these two kinds of features are combined to make a distinction between real-scene images and the corresponding recaptured ones. With a support vector machine classifier, the experimental results show that the proposed features perform well.