Understanding benford´s law and its vulnerability in image forensics

In this paper, we attempt to shed light on Benford´s law from the viewpoint of probability theory and point out its limitation in image forensic applications. First, we consider a generalized form of Benford´s law and relate it to a random variable of a certain probability density function such as the generalized Gaussian or Laplacian function. Then, we examine the application of the generalized Benford´s law to image forensic applications and point out such an application is vulnerable to the histogram manipulation attack.