Exposing Digital Image Forgeries Based on Double Quantization Mapping Relation of DCT Coefficient

Mostly, some traces of digital tamper may be left in the image during the manipulation, which result in the inconsistency of the natural images and provide some clues for image forgeries. This paper proposes a novel double compression probability model to describe the change of DCT coefficients after double compression, and combines the Bayes theorem to express the Double Quantization (DQ) effect and locate the tampered region with the posterior?probability map. Experimental results show that the method can exactly locate the tampered region, and the accuracy improves remarkably especially when the first compression quality factor is smaller than the second one. It is also found that the method is resistant to different kinds of forgery techniques, such as manual manipulation, inpainting algorithm and Bayesian-matting.