untitled

Transcription

1 ISSN , CODEN RUXUEW Journal of Software,2017,28(7): [doi: /j.cnki.jos ] Tel: ,2, 1,2, 3, 1 1 (, ) 2 ((), ) 3 (, ) :, wan@whu.edu.cn :,.,.,,.,. Copy-Move, CMYK,, CMYK 4.,,. : ;;; : TP391 :,,,..,2017,28(7): : Xie W, Wan XX, Ye ST, Jin GN. Copy-Move forgery detection for pre-printing image using digital halftoning technology. Ruan Jian Xue Bao/Journal of Software, 2017,28(7): (in Chinese). htm Copy-Move Forgery Detection for Pre-Printing Image Using Digital Halftoning Technology XIE Wei 1,2, WAN Xiao-Xia 1,2, YE Song-Tao 3, JIN Guo-Nian 1 1 (School of Printing and Packaging, Wuhan University, Wuhan , China) 2 (China Research Center for Photographic Culture Industry (Wuhan University), Wuhan , China) 3 (The College of Information Engineering, Xiangtan University, Xiangtan , China) Abstract: Pre-Printing image is referred to as digital print proof validated by copy-editor. This type of image is stored in binary screened form after being processed by raster image processor (RIP). It has indubitable legitimacy once the digital print proof is finalized. Looking at the overall printing flow, however, pre-printing images are still at the risk of malicious tampering in the storage and transmission process. Most of existing copy-move forgery detection methods employing high dimensional feature vector have high computational complexity, thus are not applicable to color separated binary screened proofs. In this paper, a new approach for copy-move forgery detection is proposed based on digital halftoning technology. The original image is binarized via digital halftoning technique in CMYK color space, which is similar to raster image processor. Then the halftone image is tiled to overlapping blocks, and the local screen density features of halftone overlapping blocks in CMYK channels are extracted for copy-move forgery detection. Experimental results : ( ); (973)(2012CB725302); (15C1324) Foundation item: National Natural Science Foundation of China ( ); National Grand Fundamental Research 973 Program of China (2012CB725302); Foundation of Hu nan Educational Committee (15C1324) : ; : , , ; : ; jos : CNKI : :39:03,

2 1836 Journal of Software Vol.28, No.7, July 2017 demonstrate that the proposed method is able to detect copy-move forgery regions precisely, and it is robust to rotation and small scale zooming attacks to forgery regions. Key words: copy-move; forgery detection; halftoning; digital image forensics,,,.,.,,.,,(:photoshop),., 10% [1].,,.,,. () [1].,,,,,.,,.,, RGB.,,...,,.,,.,,. :, CMYK,, RGB ;,, 8bit 24bit, 0 1 ;,.,,,.,, CMYK,, CMYK 4.,,,, JPEG ,,, [1]..,.

3 : 1837,,.,,,,.,. (block-based)(keypoint-based). Copy-Move,,.Fridrich [2] (DCT),,,.Popescu [3] (PCA).Kang [4] (SVD),,.Wang [5] (DWT)(DCT) JPEG.Myna [6],,,.Ardizzone [7], 5,,.Liu [8] Hu, Hu,,JPEG.Ryu [9] --, Zernike,.,Wu (LPFT) [10] (LPFFT) [11].Davarzani [12] (multiresolution local binary patterns),, k-d,,jpeg.alsawadi [13].Cozzolino [14] PatchMatch,.,Li [15],,,,,.,( ),,.(SIFT) [1618] (SURF) [19],,,.,Chen [20] Harris,.,,,,,,,.,. 2,,

4 1838 Journal of Software Vol.28, No.7, July 2017.,.,,, : CMYK, CMYK, CMYK, CMYK,,,. 1. Fig.1 The process of pre-printing image copy-move forgery detection (pre-printing image forgery detection method using digital halftoning technology). V B ; S1: CMYK ; S2: CMYK ; S3: CMYK p ij 1 9 i1 j 1 x i 1yj1 f( x, y); S4: S3 M N, B=bb; S5: S4 N B =(M b+1)(n b+1),(5)~(8) S6: N B V B k-d Tree,, d(a,b)(9); S7: AB d(a,b), d(a,b)t(t ). 2.2,,.,,. Fig.2 Halftone method based on error diffusion 2 3 [21].,,,. 2.,.x(i,j), x (, i j),b(i,j)(i,j),u (i,j). u(, i j) x(, i j) x (, i j), e(, i j) u(, i j) b(, i j) (1)

5 : x e k (2) (, i j) ( im, jn) ( m, n) m0n1 b (, i j) 0, if u 1, if u (3) t, t=0.5. 2,k (m,n),. 3, 3(b) Lena ( 3(a)). (, i j) (, i j) t t (3) Fig.3 (a) (b) Halftone image of Lena generated by error diffusion 3 Lena :(1),;(2),; (3),, ;(4). 2.3,.,(4) (i,j) p ij. p ij 1 n in2 jn2 2 x i n2 yjn2 b( x, y) (4) n=3, Fig.4 Computing method of local dot density when n=3 4 n=3 2.4,,.,,

6 1840 Journal of Software Vol.28, No.7, July 2017,.,(mean) (variance), bb B,. : E k b b p 2 ij b i0 j0 (5),E k B k,p ij (4)(i,j). : k 1 b b1 b1 2 2 pij E (6) k i0 j0 CMYK 4, E k σ k, 8,: C C Ei i M M Ei i M B Y Y Ei i K K Ei i, B, M B , N, N 2.,,.,, k-d. B M B, M B. C M Y K C M Y K V E, E, E, E,,,, (8) B i i i i i i i i N B k-d [22] 8,,. AB d(a,b). r A B A B i i i i i i (9) i1, 1 2 d A B w E E w,i ( CMYK :C=1,M=2,Y=3,K=4);r (CMYK A 4); E i, A E AB i ;, B i i (7) B i AB i ;w i,. T, d(a,b)t,(a,b). AB v AB,, v t,, t, AB, t b. 3 3.,,,,, M N O(MN).

7 : 1841, M N, N B =(M b+1)(n b+1), E k σ k O(N B b 2 ),, CMYK O(N B b 2 8)., N B k-d O(N B log 2 N B ), k-d O(log 2 N B )., 4 2 O MN O N b 8 O N log N O log N O MN log MN. B B 2 B 2 B 2, Christlein [20] Benchmark Data. Visual Studio 2010 C++, AMD Athlon X4 3.7GHz CPU,DDR3 1600HZ 8G (precision)(recall) F1-Measure, Precision Recall : TP TP Precision, Recall TP FP TP FN 2 Precision Recall F1-measure Precision Recall,TP ;FP ;FN.F1-Measure Precision Recall,. 4.2, Benchmark Data [23] () JPEG, 5 (DCT [2] PCA [3] Zernike Moments [9] SIFT [16] MLBP [12] ),,.. (1) Benchmark Data 48 48,,,. 5 96, 1. Table 1 The experimental result comparison between proposed method and other five typical methods (%) 1 5 (%) Method Precision Recall F DCT PCA ZERNIKE MLBP SIFT ,,, (10) (11)

8 1842 Journal of Software Vol.28, No.7, July 2017 MLBP,. 5,,. (a) (b) (c) (d) (e) (f) Fig.5 Example test cases for proposed copy-move forgery detection method 5 (2),,., Benchmark Data., Benchmark Data (a) 2 (b) 4 (c) 6 (d) 8 (e) 10 (f) 20 (g) 60 (h) 180 Fig.6 Example test cases for rotation attacktion 6 7,,,,. 6, 10, 60,, 180,

9 : 1843,,., 80%~120%, 80%91%93%95%97%99%101% Fig.7 The impact of rotation attacktion in feature extraction 103%105%107%109%120%. 7,,,,. 8, 91%~109%,, 80%120%,, 91%~109%,.,,,. (a) 80% (b) 91% (c) 93% (d) 95% (e) 97% (f) 99% (g) 101% (h) 103% (i) 105% (j) 107% (k) 109% (l) 120% Fig.8 Example test cases for scale attacktion 8 (3) JPEG, JPEG., ,. 9,

10 1844 Journal of Software Vol.28, No.7, July 2017 (a) 0.02 (b) 0.04 (c) 0.06 (d) 0.08 (e) 0.10 Fig.9 Example for robustness test to Gaussian noise 9 9,,,,, ( 0.06),, 0.8,,,.,. JPEG,, JPEG, JEPG,[100,20], JPEG. 10 JPEG, JEPG, JPEG, ( 10(h)),,, JPEG,,, JPEG., JPEG. (a) 100 (b) 90 (c) 80 (d) 70 (e) 60 (f) 50 (g) 40 (h) 20 Fig.10 Example for robustness test to JPEG-compression 10 JPEG , P-Time, F-Time, M-Time, O-Time. 2 5.

11 : 1845 Table 2 2 The comparison of average computation times per image in seconds 5 (:) Method P-Time F-Time M-Time O-Time DCT PCA ZERNIKE MLBP SIFT ,. 4, DCT 256, ; PCA PCA,,; ZERNIKE,; MLBP 12, LBP, LBP u2 8,1 LBP u2 12,2 LBP riu 2 16, , k-d,,. CMYK 8,,,.,, (DCTPCAZERNIKEMLBP),(SIFT),,,. 5,,, CMYK,, CMYK, k-d.,.,,,jpeg,. References: [1] Birajdar GK, Mankar VH. Digital image forgery detection using passive techniques: A survey. Digital Investigation, 2013,10(3): [doi: /j.diin ] [2] Fridrich AJ, Soukal BD, Lukáš AJ. Detection of copy-move forgery in digital images. In: Proc. of the Digital Forensic Research Workshop. Cleveland: DFRWS, [3] Popescu AC, Farid H. Exposing digital forgeries by detecting duplicated image regions. Technical Report, TR , Department of Computer Science, Dartmouth College, [4] Kang XB, Wei SM. Identifying tampered regions using singular value decomposition in digital image forensics. In: Proc. of the 2008 Int l Conf. on IEEE Computer Science and Software Engineering. IEEE, 2008,3: [doi: /CSSE ] [5] Wang XF, Zhang XN, Li Z, Wang SP. A DWT-DCT based passive forensics method for copy-move attacks. In: Proc. of the 3rd Int l Conf. on IEEE Multimedia Information Networking and Security (MINES). IEEE, [doi: /MINES ] [6] Myrna AN, Venkateshmurthy MG, Patil CG. Detection of region duplication forgery in digital images using wavelets and log-polar mapping. In: Proc. of the 2007 Int l Conf. on IEEE Computational Intelligence and Multimedia Applications. IEEE, 2007,3: [doi: /ICCIMA ]

12 1846 Journal of Software Vol.28, No.7, July 2017 [7] Ardizzone E, Bruno A, Mazzola G. Copy-Move forgery detection via texture description. In: Proc. of the 2nd ACM Workshop on Multimedia in Forensics, Security and Intelligence. ACM, [doi: / ] [8] Liu GJ, Wang JW, Lian SG, Wang ZQ. A passive image authentication scheme for detecting region-duplication forgery with rotation. Journal of Network and Computer Applications, 2011,34(5): [doi: /j.jnca ] [9] Ryu SJ, Lee MJ, Lee HK. Detection of copy-rotate-move forgery using Zernike moments. In: Proc. of the Information Hiding. Berlin, Heidelberg: Springer-Verlag, [doi: / _5] [10] Wu Q, Wang S, Zhang X. Detection of image region-duplication with rotation and scaling tolerance. In: Computational Collective Intelligence, Technologies and Applications. Berlin, Heidelberg: Springer-Verlag, [doi: / _11] [11] Wu Q, Wang S, Zhang X. Log-Polar based scheme for revealing duplicated regions in digital images. Signal Processing Letters, 2011,18(10): [doi: /LSP ] [12] Davarzani R, Yaghmaie K, Mozaffari S, Meysam T. Copy-Move forgery detection using multiresolution local binary patterns. Forensic Science Int l, 2013,231(1): [doi: /j.forsciint ] [13] AlSawadi M, Muhammad G, Hussain M, Bebis G. Copy-Move image forgery detection using local binary pattern and neighborhood clustering. In: Proc. of the Modelling Symp. (EMS), 2013 European. IEEE, [doi: /EMS ] [14] Cozzolino D, Poggi G, Verdoliva L. Copy-Move forgery detection based on patchmatch. In: Proc. of the 2014 IEEE Int l Conf. on.image Processing (ICIP). IEEE, [doi: /ICIP ] [15] Li J, Li XL, Yang B, Sun XM. Segmentation-Based image copy-move forgery detection scheme. IEEE Trans. on Information Forensics and Security, 2015,10(3): [doi: /TIFS ] [16] Amerini I, Ballan L, Caldelli R, Bimbo AD. A sift-based forensic method for copy-move attack detection and transformation recovery. IEEE Trans. on Information Forensics and Security, 2011,6(3): [doi: /TIFS ] [17] Amerini I, Ballan L, Caldelli R, Bimbo AD, Tongo LD, Serra G. Copy-Move forgery detection and localization by means of robust clustering with J-linkage. Signal Processing: Image Communication, 2013,28(6): [doi: /j.image ] [18] Costanzo A, Amerini I, Caldelli R, Barni M. Forensic analysis of SIFT keypoint removal and injection. IEEE Trans. on Information Forensics and Security, 2014,9(9): [doi: /TIFS ] [19] Shivakumar BL, Baboo LDSS. Detection of region duplication forgery in digital images using SURF. IJCSI Int l Journal of Computer Science Issues, 2011,8(4): [20] Chen LK, Lu W, Ni JQ, Sun W, Huang JW. Region duplication detection based on Harris corner points and step sector statistics. Journal of Visual Communication and Image Representation, 2013,24(3): [doi: /j.jvcir ] [21] Shi L. Research on digital halftoning technique [MS. Thesis]. Xi an: Xidian University, 2007 (in Chinese with English abstract). [22] Silpa-Anan C, Hartley R. Optimised KD-trees for fast image descriptor matching. In: Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition, CVPR IEEE, [doi: /CVPR ] [23] Christlein V, Riess C, Jordan J, Riess C, Angelopoulou E. An evaluation of popular copy-move forgery detection approaches. IEEE Trans. on Information Forensics and Security, 2012,7(6): [doi: /TIFS ] : [21].[].:,2007. (1985),,,,,,. (1983),,,,,,. (1973),,,,,,. (1973),,,,