Forevidizer: forensic video & image analyzing toolkit

Abstract

Digital videos and images have become a common thing in life. More and more sophisticated tools are becoming available for the general consumers. With the advancement of digital image processing and video processing technologies, various kinds of images and videos are produced from different perspectives. As a result videos can be used for various frauds and illegal activities. Legislative changes have been made to accept videos and images from digital cameras as witnesses for legal proceedings. Consequently there is a growing interest in forensic analysis of video content where the integrity of digital images and videos need to be checked. In this respect it has become essential to have a proper toolkit to analyze whether a particular video is a real one or one that has been tampered with. As video editing techniques are getting very sophisticated, tampered videos are hard to detect. However, when a video is tampered with, some of the basic properties of the video are changed. Then to detect those changes it is needed to use complex image processing and video processing techniques and algorithms. We present methods to analyze these properties of a given video, and produce statistical details for the video to ascertain whether it is tampered with or not, and if it is tampered with then what changes have been made. Video frame duplicate detection, video double MPEG compression detection, image double JPEG compression detection and duplicated regions within image detection are the basic methods of forgery detections.