Image splicing detection using multi-resolution histogram

Abstract

In this paper, we investigate the prospect of using multi-resolution histogram of an image as features to detect image splicing which is the operations that copy and move a part of an image to another destination. Multi-resolution histogram is a kind of statistical feature which contains image's spatial information. Then Support Vector Machine (SVM) is used to train and test as classifier for estimating whether a given digital image is a splicing image or an authentic image. The experiment shows that the method is simplicity and efficiency. © 2009 Springer-Verlag Berlin Heidelberg.