Fast and Robust Image Cloning Detection using Block Characteristics of DWT Coefficients
Michael Zimba, Xingming Sun · International Journal of Digital Content Technology and its Applications · 2011
In this paper, we present a fast and robust algorithm for detecting a specific type of image fakery, namely, cloning forgery. Discrete Wavelet Transform of a suspicious image is performed and only its low frequency sub-band is extracted in order to reduce the size of the image. A fixed window is then slid over the sub-band, pixel-by-pixel, and a feature vector is extracted at each location. The processed feature vectors, considered as rows of a matrix, are sorted using radix sort. Shift vector, which is the difference between positions associated with every pair of adjacent rows of the sorted matrix, is calculated. The frequency of each shift vector is known. Higher frequencies suggest possibility of segment duplications. Consequently, the associated blocks are mapped for further decision making. Finally, the image is morphologically processed to minimize false detections. Since the dimensions of both the image and the extracted feature vectors are reduced, the proposed algorithm is faster compared with the existing related algorithms. At the same time, experimental results show that the accuracy of detection is good.