Passive Image Forensics Based On Fourier Spectrum Analysis

Over the past years, digital images have been widely used in the Internet and many other applications. As image processing techniques are developing at a rapid speed, tampering digital images without leaving any obvious traces becomes easier and easier. Passive image forensics, a technology of detecting image authenticity and source without relying on any pre-extraction or pre-embedded information, has become a hot topic with broad prospect in multimedia security.In this dissertation, we propose several new methods to push forward the development of passive image forensics tools in terms of the Fourier spectrum of image scaling, rotation and JPEG compression.Motivated by Gallagher's image rescaling detection/estimation method, we propose an improved algorithm that can blindly estimate the scaling ratio of zooming-in or zooming-out image. Experimental results indicate the robustness of the proposed estimator against post-processing.We develop an image rotation angle estimator based on the relations between the rotation angle and the frequencies at which peaks occur in the spectrum of the image's edge map due to interpolation. Then, we use techniques of rescaling/rotation detection and parameter estimation to detect fake objects inserted into images.When a forged image contains areas from different sources, or from another part of the same image, rescaling and/or rotation are often involved. In those cases, interpolation is a necessary step. By dividing the image into blocks, detecting traces of rescaling and rotation in each block, and estimating the parameters, we can effectively expose the forged areas in rescaled or rotated images. If multiple geometrical transformations are involved, different processing sequences, including repeated zooming, repeated rotation, rotation-zooming, and zooming-rotation, can be determined from different manipulations of the peaks due to rescaling and rotation. This may provide a useful clue to image authentication.To commit a digital forgery, it is often necessary to combine together several images of different JPEG compression quality. We propose a passive approach to detect digital forgeries by checking inconsistencies of blocking artifacts. After calculating the second-order difference edge map by convolving the blocky image with a Laplacian kernel, we compute the mean of the results along the horizontal (vertical) direction and obtain a vector. Discrete Fourier transform of this vector produces a normalized spectrum curve and constructs a blocking artifact measure. The next step is to divide the composite image into overlapped blocks and calculate the measure of each block. The forged areas can be effectively exposed by checking inconsistencies of blocking artifacts. Simulation results show the effectiveness and veracity of the proposed method.