Robust Contrast Enhancement Forensics Using Convolutional Neural Networks.
Pengpeng Yang, Rongrong Ni, Yao Zhao, Gang Cao, Haorui Wu, Wei Zhao · arXiv (Cornell University) · 2018
Contrast enhancement(CE) forensics has always been of concern to image forensics community. It can provide an effective rule for recovering image history and identifying tampered images. Although a number of algorithms have been proposed, the robustness of CE forensics techniques for common pre/post processing is unsatisfactory. Attenuate such deficiency, in this letter we first present a theoretical analysis of the robustness and stability of feature space in pixel and histogram domains. Then, two kinds of end-to-end methods based on convolutional neural networks(P-CNN, H-CNN) are proposed to achieve robust CE forensics when JPEG compression and the histogram-based anti-forensics attack are applied as pre/post processing, respectively. Experimental results show that the proposed methods produce much better performance than the state-of-the-art schemes.