Forensic Detection of Image Tampering Using Intrinsic Statistical Fingerprints in Histograms
Matthew Christopher Stamm, K. J. Ray Liu · Hokkaido University Collection of Scholarly and Academic Papers (Hokkaido University) · 2009
Abstract—As the use of digital images has become more common throughout society, both the means and the incentive to create digitally forged images has increased. Accordingly, there is a great need for methods by which digital image alterations can be identified. In this paper, we propose several techniques for identifying digital forgeries by detecting the unique statistical fingerprints that certain image altering operations leave behind in an image’s pixel value histogram. Specifically, we propose methods to detect the global and local application of contrast enhancement and to detect the addition of noise to a previously JPEG compressed image. These methods are tested through a number of experiments, and results showing the effectiveness of these algorithms are discussed. I.