Blind Image Forgery Detection by using DCT and SURF Based Algorithm

Abstract

A remarkable array of visual images surrounds us. Image forgery by using digital technology has eroded confidence in the integrity of imagery. Image forensics is an aid to re-establish the trust and originality of imagery. It is important to identify the type of forgery for application of relevant image forensic technique for optimal output. Based on the availability of the original image, image forensic algorithms are classified in the blind and non-blind algorithms. But the availability of original imagery is not always necessary for all the cases of image tampering. The blind forgery detection is particularly difficult. Key point scheme is introduced instead of block based Discrete Cosine Transformation (DCT) and after the validation, the state of the art blind image forgery techniques, i.e., DCT and proposed Speeded-Up Robust-Features (SURF) algorithms are evaluated. Average Accuracy, True Positive Rate (TPR), and True Negative Rate (TNR) are used as parameters for performance evaluation on various images. Experimental results concluded the efficiency of SURF scheme in dealing with region rotation manipulations. The paper concluded with the efficiency of the key-point based scheme.