Copy-move image forgery detection using cuckoo search

Abstract

Most people face the dilemma of accepting photographs as authentic or not, mainly in the case of forensics, where the images will influence judgment. Research communities are constantly providing methods to identify these kinds of forged images. Attention capture is specifically the cases where a region of the image is copied in the same image (copy-move forgery - CMF). To detect these forged images, a recent approach named scale-invariant features transform (SIFT) has proved its worth and robustness in various geometrical transformations. However, the framework needs to be optimised for numerous parameters involved because a wrong selection of values leads to wrong identifications. To solve this problem, a novel method has been proposed named as cuckoo search-based copy-move forgery detection (CSCMFD) for optimising the parameter values in the SIFT framework. The CSCMFD demonstrates to attain better results by automatically determining the values of different parameters as compared to state of art research work. Experimentation is performed on the Christlein et al. database and MICC-F220 dataset. The experimental results proved that CSCMFD is able to capture small-forged areas as well as regions that are difficult to identify by other methods.