Blind Tampering Identification by Incorporating EFR and Calibration

Currently, images could be created and tampered easily with low cost image devices and sophisticated editing software. One cannot trust any image to be integrated anymore. Thus there is a strong need to investigate image tampering identification. This paper focuses on blind scheme of tampering identification which is practical in most scenarios. Given an image, our idea is to find its reference image rather than estimating its original image. By further analyzing empirical frequency response (EFR) incorporating calibration, the proposed blind scheme can identify several types of tampering operations. Experiment results show that the novel method classifies tampering operations accurately and also works when including integrity images.