Digital Image Forgery Detection based on Texture Feature and Clustering Technique

Abstract

The image forgery detection is important tools in digital multi-media analysis. Now a day's digital multi-media faced a problem of copy paste and tampering by different multi-media authoring tools. The tampered and copy paste image change the actual scenario of original image and its illegal process in current scenario of multi-media. For the detection of image forgery various pixel and transform based method are applied. The applied method is better in some detection and estimation, but faced a certain limitation. In this paper proposed texture based image forgery detection. The texture based image forgery detection is very efficient in terms of detection ratio. For the extraction of texture feature used discrete wavelet transform function. For the generation of block used partition clustering technique. the partition clustering technique creates the block of original and forged image. The proposed algorithm is simulated in MATLAB software and used very famous dataset MFIC2000.

An Approach to Detect Image Forgery by Discrete Wavelet Decomposition

“…However, these methods have some demerits such as over fitting and localizing the forged area. Recently, Ujjainiya et al [10] have proposed the texture feature and clustering for digital forgery detection. However, this method requires both original and forged images.…”