ANTI - FORENSICS OF MEDIAN FILTERED IMAGES USING NON - LINEAR OPTIMIZATION TECHNIQUES

Abstract

Abstract Digital image forensics is fast evolving field with many applications as the digitization is growing. Digital images have become more prone to tampering and forgery. Median filtering detection emerged as widely used tool against tampering in images. Recently a wide variety of techniques have been evolved for median filtering detection in digital images. This median filtering detection is done in blind fashion as the investigator is unaware of the post processing done on the image. Detection of median filtering is an important task in image forensics, since this operator is frequently used both for benign and malicious processing. It is also seen that median filtering detection is also used to hide traces from images so that preprocessing steps doesn’t leave any footprints on the image. This is why counter forensics is evolved. In the proposed method particle swarm optimization is applied to minimize the function derived from statistics of detection of median filtering in images in the Yuan method. Yuan method is based on calculation of feature set comprised of five discriminate features used for detection of median filtering. Particle swarm optimization is a meta-heuristic tool used for the minimization of the objective function of the characteristic feature set.