Recognizing image splicing forgeries using histogram features

Today digital images have possessed the position of the first simple analog photographs. The forgery of such images has turned out to be logically non-identifiable to the naked eyes and it is simple to do with the assistance of innovation of digital technology. Image splicing is a frequently utilized technique among the various kinds of image forgeries. So, this paper intends to offer a novel method to reveal the image splicing forgery by extracting the histogram features of Local Binary Pattern (LBP) code. These features are given to the classifiers such as combined k-NN, back propagation neural network (BPNN), and support vector machine (SVM) to recognize the spliced images along with the authentic images. The experiment is conducted with the publically accessible dataset of Columbia image splicing forgery. The performances are evaluated and their results are also compared to the previous methods.