CBFD: Coherence Based Forgery Detection technique in image forensics analysis

Abstract

In the recent times various malicious attacks can be performed over images for duplicating the images which introduces huge amount of challenges in the area of forgery detection and feature based image authentication. Researchers have focused on the development of efficient image forgery detection techniques which can be applicable to optimize image retouching attacks. The current research trends also focus on the authenticity of an image. In this paper a Coherence Based Forgery Detection (CBFD) technique has been proposed. The proposed method divides the input image into segmented blocks and extracts feature vectors from the statistically sorted matrix. The proposed methods also uses feature vectors and the distance between feature vectors for ensuring the suspicious blocks present in the image. The experimental analysis shows that the proposed system achieves better accuracy in detection of forged regions even if it is blurred with the use of Gaussian blurring which hides the forged and suspicious blocks of an image.