Copy-move Forgery Detection using DWT

Abstract

In recent years tools for editing have become powerful and common, they are also easily available, and usage of such tools has made image tampering also very common. Now-a-days it is very common to alter the image by removing or adding certain information with the images often with the objective of hiding an object or structure or either duplicating a region or a part in the image. The approach focuses on the detection of copy-move type of forgery, which is a blind passive forensic approach for forgery detection. The copy-move forgeries detection has been widely researched topic these days and in the field of blind image forensics it is one of the frequently searched topics. To make an image forged one with this technique one of the parts of original digital image is copied and pasted somewhere else to some other part in the same original image. Here, a method has been proposed which is comparatively efficient and reliable as well when compared to previous methods like DWT with phase co-relation, DCT, PCA, KPCA and others. Blind image forensics approach is described here with the purpose of detecting copy-move forgery. The proposed technique reduces forged image in terms of dimension using DWT (Discrete Wavelet Transform). In the proposed approach the image has been divided into blocks at different levels, after which feature are extracted and then sorting is performed on these. Based on threshold, the various features are compared and corresponding suspected sub-block is obtained as forged region in the input image. This method works efficiently for forged images which has undergone copy-move type of forgery.