An Efficient Method for Copy-Move Forgery Detection HWEI-JEN LIN 1 , CHUN-WEI WANG 1 and YANG-TA KAO 2 1 Department of Information Engineering and Computer Science Tamkang University 2 Department of Network Information Technology Chihlee Institute of Technology 151 Ying-Chuan Road, Tamsui Taipei, Taiwan 1 086204@mail.tku.edu.tw , http://pria.cs.tku.edu.tw 2 ydkao@mail.chihlee.edu.tw , http://int.chihlee.edu.tw Abstract: - This paper proposes a method for detecting copy-move forgery over images tampered by copying some regions and pasting them onto other regions. To detect those forgeries, we divide the given image into overlapping blocks, extract feature for each blocks, and sort the extracted feature vectors by radix sort. The difference (shift vector) of the coordinates of every pair of adjacent vectors in the sorting list is computed. The accumulated number of each of the shift vectors is evaluated. Finally, the medium filtering and connected component analysis are performed to obtain the detected result. Compared with other methods, employing the radix sort makes the detection much more efficient without degradation of detection quality. Key-Words: - forgery detection, copy-move, radix sort, connected component analysis, medium filtering. 1 Introduction With the development of Internet and image processing techniques, images can be easily acquired through internet and tampered using some commonly available software, such as Photoshop and Photoimpact. As shown in Fig. 1, the forged image “Fonda Speak To Veitnam Veterans At Anti-War Rally” in Fig. 1(a) was synthesized using the images shown in Fig.s 1(b) and 1(c). For protecting the copyright and preventing forgery with a bad intention, methods for forgery detection have become more and more important. Recently, many methods for detecting forged images have been proposed. Popescu [8] detected forgeries with linear interpolation, scaling or rotation based on the relationship between each pixel and its neighbors. Nillius et al. [2] and Johnson et al. [7] used light source consistency to detect forged images. Li et al. [10] detected tampered watermarked images with the embedded information and then recover the images. Defects of cameras such as chromatic aberration and sensor pattern noise, and the color filter arrays the cameras use for interpolating colors can be used to detect forgeries [4][9][12]. Copy-move images are easily made by copying certain regions and pasting them on some other regions. The existing methods [1][5][6][11] for detecting such kind of forgeries are all time- consuming. In this paper, we shall propose a simple and fast method to detect copy-move forgeries. Compared with other methods, our algorithm is more efficient without degradation of detection rates. The rest of this paper is organized as follows. Related work is discussed in Section 2. In section 3, the proposed method is described in details. In Sections 4 and 5, we show some experimental results and make a conclusion for this paper. (a) (b) (c) Fig. 1 (a). A synthesized image “Fonda Speak To Veitnam Veterans At Anti-War Rally”, (b)&(c). original images. Proceedings of the 8th WSEAS International Conference on Applied Computer and Applied Computational Science ISSN: 1790-5117 250 ISBN: 978-960-474-075-8