M.Sridevi, C.Mala and S.Sandeep, National Institute of Technology, India

Image forgery is the manipulation of digital images to conceal meaningful information or objects in the image. Among different image forgery techniques, copy – move forgery is one of the frequently used passive image forgery approach. The existing methods such as Principle Component Analysis (PCA), Discrete Wavelet Transform (DWT) & Singular Value Decomposition (SVD) are time consuming. Hence it is not suited for digital forensic science, surveillance system applications which uses image, video or multimedia security. This paper proposes a parallel algorithm for the copy – move image forgery detection to decrease execution time of the algorithm. The method uses overlapping blocks and lexicographical sorting in a parallel manner. The simulation results show that the proposed parallel version detects the forged region faster, so that it is best suited for real time applications.

Copy-move detection, Image forgery, Block matching, Parallel environment, Complexity.