Develop an algorithm for image forensics using feature comparison and sharpness estimation

The paper presents a new method to detect the forgery in Copy-Move images using feature comparison to find the similar parts and sharpness estimation to collect the suspicious edges. Besides, one-level DWT decomposition, with the role of multiresolution, is used to limit the computational complexity and morphological operation is applied for presenting counterfeit objects. Both feature comparison and sharpness are done in overlapping 16×16 blocks of LL and HH sub-bands. Searching similar blocks in approximation part by modified Zernike Moments and calculating the sharpness of edges in HH sub-band confirm the copy-move objects in image. Faked regions have features similar to any other image's region and sharpness at boundaries higher than at other edges in the image. The combination of LL and HH sub-band can improve the reliability of the detection. The suspicious regions are then colored by opening and closing of morphological analysis and shown by logical operator "AND" to the tested image. The fact that similar regions can be detected and defined to be copy-move or same texture is the novelty of the proposed method. Results are simulated in Matlab environment with high accuracy and also compared to some related methods.