J. of Advanced Engineering and Technology Vol. 1, No. 1 (2008) pp. 45-51 f m s p» w Á xá zá Ÿ Á w m œw Image Retrieval Based on Gray Scale Histogram Refinement and Horizontal Edge Features Sang-Uk Shin, Hui Zhao, Jung-Il Hoe, Gwang-Won Kang, and Jong-An Park Information Communication Engineering, Chosun University, Korea (Received : Aug. 28, 2008, Revised : Sep. 04, 2008, Accepted : Sep. 12, 2008) Abstract : This study realized a retrieval system which uses gray scale histogram and edge information to decrease errors occurring in simple keyword input system of existing retrieval systems. This retrieval system composes of the following five steps: abstracting features of images, abstracting image refinement and edge information, securing necessary information from analysed features, retrieving secured information from database, and comparing and abstracting images from retrieved databases. The proposed retrieval system aims at quick and exact retrieval, which is to be demonstrated through simulation. Keyword : histogram refinement, gray scale, horizontal edge, CBIR. 1. ful l w w, v,, p l w ƒwš. p ƒ w l w. z,,» v w š,» w ƒ y w š [1-4]. l»» (text-based image retrieval) ü» (content-based image retrieval) [5].» ü ƒ wš l l w ƒ w».» x wš ƒ. ù w w w. ü» ü e (color), (texture), (shape) ƒ p ü wš l l w p Corresponding Author Tel : 062-230-7064 E-mail : japark@chosun.ac.kr w».»» w ƒ w w ƒ w w ƒ w š.. 2 f š g š w wš, 3 w f m w yw s w l w w. š 4 x š, 5. 2. s w s,,ƒ yw p y ü,, w. w m m j», s ƒ w y z w. v e œ w» w, y bin w v s ƒ w. e m f yw m w w š 45
46 Á xá zá Ÿ Á w w w. 2.1.» š ( f š ) û w yw w yw w» w f y w w. Figure 1. f ywš š s j y e w. š s 4 ù z 4 ƒ tx 4 z w. w ƒ 4 4 w 16 p ƒ š p l w l ü ƒ p w w. yw ƒ. ù m 4 w w w l w š ƒ ù š s z. 6CDNG'ZKUVKPI*KUVQITCO4GHKPGOGPV#NIQTKVJO(GCVWTG6CDNG Sub-divisions 1 2 3 4 Bin 1 1181 1019 834 820 Bin 2 782 610 582 585 Bin 3 554 510 432 730 Bin 4 452 672 584 549 2.2.» š ( g š ),,ƒ yw p y ü,, w g w w š. Figure 2. y wš yw w w g z w. e yw g z w. 3 3v ww v ww p l ü ƒ p w w [11-12]. g š l yw wù ý ƒ š. ù g ü 3 3 v w w g w l ƒ» v w ù. w š p l l ù l. 3. w š w š» š m ã y w ¾ w. Figure 3 w š. Figure 4 f y wš (KIWTG'ZKUVKPIJKUVQITCOTGHKPGOGPVCNIQTKVJO œw» 1«1y (2008)
f m s p» w 47 (KIWTG%QTPGTKPHQTOCVKQPGZVTCEVKQPCNIQTKVJOQHGZKUVKPIGFIGU y» w m y w. y e» 16 ù. 16 ù Figure 5 y w. yw ƒƒ 16 Figure 6 (KIWTG2TQRQUGFCNIQTKVJO s w. s w ƒƒ s ƒ š p l w. p ƒ š l w z w. J. of Adv. Eng. and Tech., Vol. 1, No. 1 (2008)
48 Á xá zá Ÿ Á (KIWTG)TC[UECNGKOCIGWPKHQTOK\CVKQP (KIWTG'FIGKOCIGUGZVTCEVGFHTQOGCEJ$KP (KIWTG+OCIGVQCESWKTGJQTK\QPVCNKPHQTOCVKQPQHGFIGU 4. x ü» e j l ü ƒ w. w š w» w l p Figure 8, (1),(6),(11) w Figure 8, (2,3,4,5,), (7,8,9,10), (12,13,14,15) ù y wš w w. ƒƒ w p l l š w w. Figure 9» f š. z 5 w. 5 2 z w. œw» 1«1y (2008)
f m s p» w 49 (KIWTG1TICPK\CVKQPQHHGCVWTGVCDNG (KIWTG6GUVKOCIGU (KIWTG4GUWNVUQHITC[UECNGTGHKPGOGPVCNIQTKVJOTGVTKGXC J. of Adv. Eng. and Tech., Vol. 1, No. 1 (2008)
50 Á xá zá Ÿ Á w œ. Figure 10» g š. z 5 w. ù w y w. 4 3 z w. w z. Figure 11 w s z š.» š w w š ý. w, w š» š yw w ü x mw w. Table 2» š w» w Recall Precision w. Recall l (KIWTG4GVTKGXCNTGUWNVUQHEQTPGTKPHQTOCVKQPGZVTCEVKQPCNIQTKVJOQHGZKUVKPIGFIGU (KIWTG4GUWNVUQHTGVTKGXCND[RTQRQUGFGFIGJQTK\QPVCNKPHQTOCVKQPCESWKUKVKQPKOCIGTGVTKGXCNCNIQTKVJO š 6CDNG%QORCTCVKXG4GVTKGXCNQH'ZKUVKPI#NIQTKVJOYKVJ2TQRQUGF#NIQTKVJO» š š g f s + s 32.647350 seconds 7.647350 seconds 1.507928 seconds 1.504574 seconds 1.895535 seconds Recall 0.3275 0.4285 0.5714 0.7142 0.7343 Precision 0.3965 0.5128 0.6124 0.6927 0.7133 œw» 1«1y (2008)
f m s p» w 51»y w ( 1) ù ký. Rr Recall = ----- T ( 1) Precision l» y w ( 2) ùký. Rr Precision = ----- Ri ( 2) Recall Precision [0, 1] ƒ, 1 ƒ ¾ ƒ w. 5. ü»»» f m ƒ š y w s w w l w.» f š g š w j» w s z w l p l w w. s p w» ƒ œ ƒ š w. w z w» w» f m š g š w w w y w. w š p w l» f š 3.542%, g š 20.8% l ƒ y w. w š» f š 6.14 seconds, g š 31.143 second w. yw š w ƒ. š» f š 0.18%, g š 0.28% yw ƒ x mw w. y Ÿ w ygld yg ld» (CT) w. š x (1) Smith, J. R., and Chang, S. F., Transform features for texture classification and discrimination in large image databases, Proc IEEE conf. Image (1994). (2) Feng, D., Siu, W. C. and Zhang, H. J., Fundamentals of Content-based Image retrieval, in Multi media Information Retrieval and Management-Technological Fundamentals and Applications, New York, NY, Springer (2003). (3) Gonzalez, R. C. and Woods, R. E., Digital Image Processing, Prentice Hall (2002). (4) Stricker, M. A. and Dimai, A., Color indexing with weak spatial constraints, In Storage and Retrieval for Still Image and Video Databases, 2670, 29-40 (1996). (5) Mehtre, B. M., Kankanhalli, M. S. and Lee, W. F. Content- Based Image Retrieval Using A Composite Colo r-shape Approach, Information Processing & Management, 34(1), 109-120 (1998). (6) Swain, M. J. and Ballard, D. H. Color Indexing, Int. J. Com. Vis., 7(1), 11-321 (1991). (7) Funtand, B.V., Finlayson, G. D., Color constant color indexing, IEEE Trans. on Pattern Analysis and Machine Intelligence, 17, 522-529 (1995). (8) Vertan, C. and Boujemaa, N., Spatially constrained color distributions for image indexing, International Conf erence on Color in Graphics and Image Processing -CGIP'2000, INRIA Rocquencourt, France (2000). (9) Chun, J. C., Digital Image Processing, 1, 73-82 (2005). (10) Comaniciu, D., Meer, P., Xu, K., and Tyler, D., Retrieval performance improvement through low rank corrections, Proc IEEE conf CBAIVL '99, 50-54 (1999). (11) Nishat Ahmad, Jongan Park, Corner geometry representation using code vectors for image retrieval, The 7th IEEE International Symposium on Signal Processing and Information Technology (ISSPIT 2007), Cairo, Egypt (2007). (12) Nishat Ahmad, Jongan Park, Corner geometry representation using code vectors for image retrieval, The Second IEEE International Conference on Digital Information Man agement (ICDIM 2007), Lyon, FRANCE (2007). J. of Adv. Eng. and Tech., Vol. 1, No. 1 (2008)