A robust technique for splicing detection in tampered blurred images

Image splicing is a common way of image forgery wherein a part of an image is cropped and pasted to another image. The rapid growth of various image editing softwares has led to the generation of a large number of forged images. Image splicing detection can be made possible by the analysis of a collection of features or characteristics of an image which distinguishes it from an authentic image. Splicing detection can be more challenging if the forger uses some post-processing operations after splicing. After a normal cut and paste of a portion of an image, the composite image so formed can be made to undergo other post-processing operations like blurring of edges, resizing, rotation, noise addition etc. to remove the traces of forgery. The objective of this paper is to develop a technique for splicing detection in case of tampered blurred images in which the original and spliced regions have different blur types. The inconsistencies in the blur types of tampered images are considered for the detection of splicing in these images. The proposed method is also robust to noise attacks.