Digital image forensices

Abstract

The unique stature of photographs as a definitive recording of events is being diminished due, in part, to the ease with which digital images can be manipulated and altered. Although good forgeries may leave no visual clues of having been tampered with, they may, nevertheless, alter the underlying statistics of an image. In this project, the author tried to justify whether the given image sets were up sampled or down sampled at certain resampling rate by two approaches and therefore determine which image was closer to the original one. One was to detect the periodic pattern in the probability map across images using Expectation/Maximization algorithm. The other was to resample the smaller one to the same size of the larger image and then analyze the difference by Gabor filter and produce the classification by Fisher Linear Discriminant Analysis.