Luận văn thạc sĩ tìm kiếm ảnh trong tập dữ liệu ảnh lớn dựa trên đặc trưng

B TR NGă GIÁO D CăVÀă ÀOăT O I H C CÔNG NGH TP HCM - LÊ CƠNG KHANH TÌM KI M NH TRONG T P D L N D AăTRểNă LI U NH CăTR NG LU NăV NăTH CăS Chuyên ngành : CÔNG NGH THÔNG TIN Mã s ngành: 60480201 TP H CHÍ MINH, thángă11ăn mă2016 B TR NGă GIÁO D CăVÀă ÀOăT O I H C CÔNG NGH TP HCM - \ LÊ CƠNG KHANH TÌM KI M NH TRONG T P D L N D AăTRểNă LI U NH CăTR NG LU NăV NăTH CăS Chuyên ngành : CÔNG NGH THÔNG TIN Mã s ngành: 60480201 CÁN B H NG D N KHOA H C: TS NGUY N THANH BÌNH TP H CHệăMINH,ăthángă11ăn mă2016 CƠNG TRÌNH TR Cán b h NG C HỒN THÀNH T I I H C CƠNG NGH TP HCM ng d n khoa h c: TS.ăNGUY NăTHANHăBỊNH (Ghi rõ h , tên, h c hàm, h c v ch ký) Lu n v n Th c s đ c b o v t i Tr vào ngày 17 tháng 12 n m 2016 ng i h c Công ngh TP HCM Thành ph n H i đ ng đánh giá Lu n v n Th c s g m: (Ghi rõ h , tên, h c hàm, h c v c a H i đ ng ch m b o v Lu n v n Th c s ) TT đ H tên Ch c danh H i đ ng Ch T ch Ph n bi n Ph n bi n y viên y viên, Th ký Xác nh n c a Ch t ch H i đ ng đánh giá Lu n v n sau Lu n v n đư c s a ch a (n u có) Ch t ch H i đ ng đánh giá Lu n v n TR NG H CÔNG NGH TP HCM C NG HÒA XÃ H I CH NGH A VI T NAM PHÒNG QLKH ậ TS H c l p ậ T ậ H nh phúc TP HCM, ngày … tháng… n m 2016 NHI M V LU NăV NăTH CăS H tên h c viên: Lê Cơng Khanh Gi i tính: Nam NgƠy, tháng, n m sinh: 17-6-1977 N i sinh: Ti n Giang Chuyên ngành: Công Ngh Thông Tin MSHV: 1441860050 I- Tên đ tài: Tìm ki m nh t p d li u nh l n d a đ c tr ng II- Nhi m v n i dung: - Nghiên c u k thu t tìm ki m nh theo n i dung, t đó, đ xu t mơ hình tìm ki m nh t p d li u nh l n d a đ c tr ng nh - Xây d ng h th ng tìm ki m nh t p d li u nh l n d a đ c tr ng nh III- Ngày giao nhi m v : 23/01/2016 IV- Ngày hoàn thành nhi m v : V- Cán b h CÁN B H ng d n: Ti n S Nguy n Thanh Bình NG D N (H tên ch ký) KHOA QU N LÝ CHUYÊN NGÀNH (H tên ch ký) i L IăCAMă OAN Tôi xin cam đoan đơy lƠ cơng trình nghiên c u c a riêng tơi Các s li u, k t qu nêu Lu n v n lƠ trung th c vƠ ch a t ng đ c công b b t k cơng trình khác Tơi xin cam đoan r ng m i s đư đ giúp đ cho vi c th c hi n Lu n v n nƠy c c m n vƠ thơng tin trích d n Lu n v n đư đ c ch rõ ngu n g c H c viên th c hi n Lu n v n Lê Công Khanh ii L IăCỄMă N Tr c h t em xin bày t lòng bi t n sơu s c nh t t i th y giáo h ng d n TS Nguy n Thanh Bình đư t n tình giúp đ em r t nhi u su t trình tìm hi u nghiên c u hồn thành báo cáo lu n v n Em xin chân thành c m n th y cô đư trang b cho em nh ng ki n th c c b n c n thi t đ em có th hồn thành lu n v n nƠy Xin g i l i c m n đ n b n bè nh ng ng i bên em đư đ ng viên t o u ki n thu n l i cho em, t n tình giúp đ ch b o em nh ng em cịn thi u sót trình làm báo cáo lu n v n Cu i em xin bày t lòng bi t n sơu s c t i nh ng ng i thơn gia đình đư giƠnh cho em s quan tơm đ c bi t vƠ đ ng viên em Vì th i gian có h n, trình đ hi u bi t c a b n thân nhi u h n ch Vì v y, đ án khơng tránh kh i nh ng thi u sót, em r t mong nh n đ c s đóng góp Ủ ki n c a t t c th y cô giáo c ng nh b n bè đ lu n v n c a em đ thi n h n Em xin chân thành c m n! Tác gi lu n v n Lê Công Khanh c hồn iii TĨM T T S t ng khơng ng ng v l ng đ ng nh Web t o ngu n nh phong phú đáp c ngu n cung nh cho nhu c u c a ng i M c dù m t s mơ hình tìm ki m nh đư đ i đáp ng ph n nhu c u tìm ki m nh, song nâng cao ch t l ng tìm ki m v n đ đ c đ t Bài tốn tìm ki m nh, nâng cao ch t l ng x p h ng nh đư vƠ nh n đ c s quan tơm đ c bi t tài c a lu n v n: ắTìm ki m nh t p d li u nh l n d a đ c tr ngắ nh m đ gi i quy t toán Nhi m v c a lu n v n lƠ nghiên c u k thu t tìm ki m nh theo n i dung vƠ đ xu t mơ hình tìm ki m nh t p d li u nh l n d a đ c tr ng nh u tiên, lu n v n nêu m t s đ c tr ng c b n c a nh c ng nh m t s nghiên c u liên quan v đ c tr ng nh Ti p theo, lu n v n đ xu t ph ng pháp tìm ki m nh theo n i dung, đ a mơ hình ph i h p đ c tr ng k t h p v i toán t LTP xây d ng gi i thu t truy v n Lu n v n đư đ c hi n th c trình truy v n nh b ng gi i thu t đư đ xu t Quá trình th c nghi m đ đ it c th c hi n nhi u t p dataset khác nhau, ch a ng khác nh ng i, xe, hoa, c nh… t p d li u c a Wang t i đ a ch http://wang.ist.psu.edu/docs/related/ T p d li u bao g m 10.000 nh th nghi m β ph ng pháp: Truy xu t b ng query truy xu t d a ph ng pháp support vector machine (SVM) Qua k t qu th c nghi m, th y r ng ph t t h n ph ng pháp SVM Ph ng pháp đ xu t cho k t qu ng pháp đ xu t đư k t h p đ c tr ng c c b toàn c c l i v i ơy lƠ lỦ gi i thích t i ph ph ng pháp SVM ng pháp đ xu t t t h n iv ABSTRACT The continuous increase in volume images on the web to create images rich source supply to meet the demand picture for the people Although some models were launched image search partly meet the need to look for images, but improve the quality of search is always a problem arises Image search problem and improve image quality ratings have been received special attention The theme of the thesis: "Content-based image retrieval in the large image database based on the characteristic" in order to solve the problem on The task of the thesis is to study the technical content image search and search suggestions model images in large image data sets based on image features Firstly, the thesis outlined some basic characteristics of the image as well as a number of studies related to the image features Secondly, the thesis proposed the method for image searching by contents, making coordination model characteristics associated with LTP and building operators query algorithms Thesis has been realized by the image retrieval algorithm was proposed The processing of this work on many different training dataset, containing various objects such as people, cars, flowers, landscape in the data set of Wang at http://wang.ist.psu.edu/docs/related/ This dataset includes 10,000 images We test on two methods: Retrieve query and retrieval using methods based on support vector machine (SVM) Through experimental results, we found that the proposed method gives better results than SVM method The method propose combining local characteristics and wikis together This is the reason why the proposed method better methods based on SVM method v M CL C L I CAM OAN i L I CÁM N ii TÓM T T iii ABSTRACT iv M C L C v DANH M C T VI T T T vii DANH M C CÁC B NG viii DANH M C CÁC BI U CH , TH , S , HÌNH NH ix NG 1: GI I THI U 1.1 Gi i thi u đ tài 1.2 M c tiêu c a đ tài 1.3 N i dung đ tài 1.4 Gi i h n c a đ tài 1.5 Ph ng pháp nghiên c u 1.6 C u trúc lu n v n CH NG β: C S LÝ THUY T VÀ CÁC NGHIÊN C U LIÊN QUAN 2.1 C s lý thuy t β.1.1 c tr ng mƠu s c .4 β.1.β c tr ng k t c u .10 β.1.γ c tr ng hình d ng 16 2.1.4 Mô t Boundary .16 2.1.5 Mô t theo Vùng (Region) .17 β.β đo t ng đ ng 19 β.β.1 đo v màu s c 19 β.β.β đo v k t c u 20 β.β.γ đo v hình d ng 21 2.3 Các nghiên c u liên quan 24 β.γ.1 Trong n c 24 β.γ.β NgoƠi n c 24 vi CH NG γ: PH NG PHÁP XU T TÌM KI M NH THEO N I DUNG.26 3.1 Yêu c u toán .26 3.2 Mô hình nghiên c u .28 3.2.1 Mơ hình ph i h p đ c tr ng nh 28 3.2.2 K t h p toán t LTP (Local Ternary Pattern) 29 γ.γ Ph ng pháp rút trích đ c tr ng nh truy v n nh 32 3.3.1 Ph i h p đ c tr ng đ truy v n nh 32 3.3.1.1 X lỦ c s d li u 32 3.3.1.2 X lý nh truy v n 32 γ.γ.1.γ o s t ng t gi a véc t nh 32 3.3.1.4 Hi n th k t qu tr v 33 3.3.2 Truy v n nh dùng moments c a LTP 33 3.3.2.1 Moments 33 3.3.2.2 M u tam phân (LTP) Moment 34 γ.γ.β.γ Ph CH ng pháp đ xu t 35 NG 4: K T QU TH C NGHI M 37 4.1 T p d li u th nghi m 38 4.2 K t qu truy v n 39 4.3 Code đ c tr ng 45 CH NG 5: K T LU N 48 5.1 K t qu đ t đ 5.β u m vƠ nh 5.β.1 c .48 c m c a gi i thu t đ xu t 48 u m c a gi i thu t đ xu t 48 5.β.β Nh c m c a gi i thu t đ xu t 48 5.3 óng góp c a lu n v n 48 5.γ.1 óng góp khoa h c 49 5.3.2 óng góp th c ti n 49 5.4 H ng m r ng .49 TÀI LI U THAM KH O 50 42 Hình 4.7 K t qu truy xu t B ng 4.1 K t qu truy xu t nh c a ph S th t nh test S nh trùng th c t Ph ng pháp đ xu t so v i ph ng pháp truy xu t d a SVM Ph ng pháp khác ng pháp đ xu t 12 12 12 10 10 10 21 21 21 14 14 14 15 13 15 5 14 12 14 3 43 6 10 2 11 12 12 12 12 10 10 10 13 21 20 21 14 14 14 14 15 15 13 15 16 5 17 14 12 14 18 3 19 6 20 2 21 2 22 1 23 12 11 12 24 15 15 15 25 11 10 11 26 15 15 15 27 11 11 11 28 9 28 3 30 1 31 12 12 12 32 7 44 33 5 34 12 12 12 35 11 11 11 36 4 37 14 12 14 38 3 39 0 40 1 41 4 42 3 43 1 44 17 17 17 45 11 12 11 46 7 47 10 10 10 48 1 49 9 50 10 11 10 Chúng th nhi u thí nghi m khác Trong khn kh lu n v n này, ch nêu 50 tr ng h p c s d li u 300 nh T k t qu b ng 4.1 nhi u k t qu khác, th y r ng ph h n ph ng pháp SVM Ph toàn c c l i v i ph ng pháp SVM ng pháp đ xu t cho k t qu t t ng pháp đ xu t đư k t h p đ c tr ng c c b ơy lƠ lỦ gi i thích t i ph ng pháp đ xu t t t h n 45 4.3 Code đ c tr ng Hàm truy xu t nh function btn_BrowseImage_Callback(hObject, eventdata, handles) [query_fname, query_pathname] = uigetfile('*.jpg; *.png; *.bmp', 'Select query image'); if (query_fname ~= 0) query_fullpath = strcat(query_pathname, query_fname); imgInfo = imfinfo(query_fullpath); [pathstr, name, ext] = fileparts(query_fullpath); % fiparts returns char type if ( strcmp(lower(ext), '.jpg') == || strcmp(lower(ext), '.png') == || strcmp(lower(ext), '.bmp') == ) queryImage = imread( fullfile( pathstr, strcat(name, ext) ) ); % Truy xu t image features queryImage = imresize(queryImage, [384 256]); if (strcmp(imgInfo.ColorType, 'truecolor') == 1) hsvHist = hsvHistogram(queryImage); autoCorrelogram = colorAutoCorrelogram(queryImage); color_moments = colorMoments(queryImage); % Scale image img = double(rgb2gray(queryImage))/255; [meanAmplitude, msEnergy] = gaborWavelet(img, 4, 6); % = number of scales, = number of orientations 46 wavelet_moments = waveletTransform(queryImage, imgInfo.ColorType); % construct the queryImage feature vector queryImageFeature = [hsvHist autoCorrelogram color_moments meanAmplitude msEnergy wavelet_moments str2num(name)]; elseif (strcmp(imgInfo.ColorType, 'grayscale') == 1) grayHist = imhist(queryImage); grayHist = grayHist/sum(grayHist); grayHist = grayHist(:)'; color_moments = [mean(mean(queryImage)) std(std(double(queryImage)))]; [meanAmplitude, msEnergy] = gaborWavelet(queryImage, 4, 6); % = number of scales, = number of orientations wavelet_moments = waveletTransform(queryImage, imgInfo.ColorType); % construct the queryImage feature vector queryImageFeature = [grayHist color_moments meanAmplitude msEnergy wavelet_moments str2num(name)]; end % update handles handles.queryImageFeature = queryImageFeature; handles.img_ext = ext; handles.folder_name = pathstr; guidata(hObject, handles); helpdlg('Proceed with the query by executing the green button!'); % Clear workspace clear('query_fname', 'query_pathname', 'query_fullpath', 'pathstr', 'name', 'ext', 'queryImage', 'hsvHist', 'autoCorrelogram', 'color_moments', 'img', 'meanAmplitude', 'msEnergy', 'wavelet_moments', 'queryImageFeature', 'imgInfo'); else 47 errordlg('You have not selected the correct file type'); end else return; end 48 CH 5.1 K t qu đ tăđ NGă5:ăK TăLU N c Qua trình th c hi n lu n v n, đư hi u đ c s d li u vƠ sau tìm ph c tính ch t c a nh ng pháp phù h p v i t ng lo i nh i v i nh nhi u ánh sáng tr i ta có th dùng hàm l c đ xóa nh ng pixel b nhi u T đem l i nh ng k t qu r t t t D a cơng trình nghiên c u vƠ ngoƠi n c kho ng th i gian g n đơy, lu n v n đư đ xu t m t gi i thu t truy v n nh d a đ c tr ng c a nh Sau trình th c nghi m nghiên c u lý thuy t, tơi đư có m t vài k t qu kh quan nh : hoƠn thi n nh ng Ủ t ng c ng nh hi n th c đ ki m ch ng th c t tƠi đư ti n hƠnh đánh giá k t qu d a ph đ nh l ng pháp đ nh tính ng, c ng nh ti n hành so sánh k t qu t ng giai đo n v i ph x lý g n đơy đư đ 5.2 ng pháp c đ xu t u m vƠănh căđi m c a gi i thu t đ xu t 5.2.1ă uăđi m c a gi i thu tăđ xu t Gi i quy t đ c toán truy v n nh d a đ c tr ng c a nh, giúp ích cho q trình tìm ki m nh K t qu thu đ c đư c i thi n đ c ph n ch t l ng tìm ki m nh 5.2.2ăNh căđi m c a gi i thu tăđ xu t M c dù gi i quy t đ c vi c tìm ki m nh, nh ng d a trình xây d ng thu t gi i c ng nh th c nghi m v n t n t i m t s v n đ sau đơy mƠ có th xem nh lƠ nh c m c a đ tài: - Gi i thu t ch m i x lý v i nh s , nh 2D - T p d li u nh th nghi m h n ch m t s nguyên nhân h n ch thu th p ngu n d li u nh m u - Ch a tính đ n t c đ x lỦ c ng nh đ ph c t p c a gi i thu t 5.3 óng góp c a lu năv n 49 5.3.1 óngăgópăkhoaăh c tƠi đư đóng góp: - Lý thuy t v lo i nh - Lý thuy t v ph - tƠi đ xu t đ ng pháp truy v n nh c m t gi i thu t k t h p đ c tr ng c c b toàn c c v i đ truy v n nh -T ođ c m t tài li u tham kh o t t cho nh ng quan tơm đ n l nh v c x lý nh 5.3.2 óngăgópăth c ti n tƠi đư ti n hành tìm hi u đ đ c tr ng c b n c a m t nh đư đ c ph ng pháp truy v n nh ph bi n Các c nêu rõ c ng nh yêu c u c b n truy v n nh, không làm m t mát thông tin c a nh 5.4 H ng m r ng Do trình thu th p d li u m u nh t nhiên g p nhi u khó kh n nên s l ng nh h n ch , đ có th kh ng đ nh k t qu gi i thu t m t cách tri t đ h n vi c b sung thêm t p d li u m u Không nh ng th , vi c b sung s l ng nh v n c n b sung thêm vi c đa d ng v th lo i nh kích c Gi i thu t v n t n t i h n ch ch gi i quy t nh nh 2D, nên v n đ đ t kh c ph c nh m giúp gi i thu t x lỦ đ h n ơy lƠ nh ng xu h c v i nh 3D nhi u kích c ng phát tri n th c t hi n Do nh ng h n ch v m t ph n c ng nên đ tƠi ch a th ti n hƠnh đánh giá đ c th i gian th c hi n 50 TÀI LI U THAM KH O [1] Stricker, M and Orengo, M (1995),‘Similarity of Color Images’, Proc of the SPIE Conf, vol 2420, pp 381ậ392, 1995 [2] Ioka, M (1989),‘A Method of Defining the Similarity of Images on the Basis of Color Information’, Tech Report RT-0030, IBM Tokyo Research Lab [γ] Vassilieva, N and Novikov, B (β005)‘Construction of Correspondences between Low-level Proc of the 7th Perspective Characteristics and Semantics of Static Images’, Scientific Conf ‘Electronic All-Russian Methods and Technologies, Electronic Libraries: Collections’ RCDL’β005, Yaroslavl’, Russia [4] Stricker, M and Dimai, A (1997), ‘Spectral Covariance and Fuzzy Regions for Image Indexing’, Machine Vision Applications, vol 10, pp 66ậ73 [5] Hong-Bo Zhang & Shang-An Li & Shu-Yuan Chen & Song Zhi Su & Der-Jyh Duh & Shao Zi Li.(β01β)‘Adaptive photograph retrieval method Multimedia Tools & Applications’, DOI 10.1007/s1104β-012-1233-7 [6] Jun Yue, Zhenbo Li, Lu Liu and Zetian Fub (β011),‘Content-based image retrieval using color and texture fused features’, Mathematical and Computer Modelling, vol.54, pp 1121ậ1127 [7] Daisy, M.M.H., Morphological TamilSelvi, S and Operations for Image Prinza, L (β01β)‘Gray Retrieval’, β01β Scale International Conference on Computing, Electronics and Electrical Technologies [ICCEET], pp 571-575 [8] Chaobing Huang, Yarong Han, Yu Zhang (β01β),‘A Method for Object-based Color Image Retrieval’, Fuzzy Systems and Knowledge Discovery (FSKD), 2012 9th International Conference on , pp:1659-1663 [9] Fernando, R and Kulkarni, S (β01β), ‘Hybrid Technique for Colour Image Classification and Efficient Retrieval based on Fuzzy Logic and Neural 51 Networks’, Neural Networks (IJCNN), The β01β International Joint Conference on, pp:1-6 [10] Zhu Qiaoqiao, Method Based Huang Yuanyuan (β01β),‘A on Color New Image Feature’, Intelligent System Retrieval Design and Engineering Application (ISDEA), 2012 Second International Conference on, pp:56-59, 2012 [11] Rasli, R.M.Muda, T.Z.T., Yusof, Y.,Bakar, J.A (β01β)‘Comparative Analysis of Content Based Image Retrieval Technique using Color Histogram A Case Study of GLCM and K-Means Clustering’, Intelligent Systems, Modelling and Simulation (ISMS), Third International Conference on, pp: 283 ậ 286 [12] Content-based binary image retrieval using the density histogram Panagiotis adaptive hierarchical Sidiropoulos, Stefanos Vrochidis, Ioannis Kompatsiaris, Pattern Recognition vol 44, pp 739ậ750 [13] Konstantinos Konstantinidis, Vasileios Vonikakis, Georgios Panitsidis and (β011),‘A Center-Surround Histogram for content-based Ioannis Andreadis image retrieval’, Pattern Anal Applic, vol 14, pp β51ậ260 [14] Xiang-Yang Wang & Jun-Feng Wu & Hong-Ying Yang(β010), ‘Robust image retrieval based on colour histogram of local feature regions’, Multimed Tools Appl vol 49, pp 323ậ345 [15] Haojie Li, Xiaohui Wang, Jinhui Tang and Chunxia Zhao(β01γ), ‘Combining global and local matching of multiple features for precise item image retrieval’, Multimedia Systems vol 19, pp 37ậ49 [16] Haralick, R.M., Shanmugam, K., and Dienstein, I.(197γ), ‘Textural Features for Image Classification’, IEEE Trans.Systems, Man Cybernetics, vol γ, no 6, pp 610ậ 621 [17] Howarth, P and Rüger, S (β004), ‘Evaluation of Texture Features for Contentbased Image Retrieval’, Proc Of CIVR'04, pp γβ6ậ334 52 [18] Jiayin Kang and Wenjuan Zhang, ‘A Framework for Image Retrieval with Hybrid Features’, Control and Decision Conference (CCDC), β01β β4th Chinese, pp: 1326 ậ 1330, 2012 [19] Yong-Hwan Lee, Sang-Burm Rhee, Bonam Kim, ‘Content-based Image Retrieval Using Wavelet Spatial-Color and Gabor Normalized Texture Multi-resolution Database’, Innovative Ubiquitous Mobile in and Internet Services in Computing (IMIS), 2012 Sixth International Conference on, pp: 371 ậ 377, 2012 [20] Smith, J.R and Chang, S.-F (1994), ‘Transform Features For Texture Classification and Discrimination in Large Image Databases’, Proc of IEEE Int Conf on Image Processing (ICIP-94), Austin [β1] Do, M.N and Vetterli, M (β000), ‘Texture Similarity Measurement Using KullbackậLeibler Distance on Wavelet Subbands’, Proc of Int Conf on Image Processing, 2000, vol 3, pp 730ậ733 [22] Sumana, I.J., Guojun Lu, Dengsheng Zhang (201β),‘Comparison of Curvelet and Wavelet Texture Features for Content Based Image Retrieval’, Multimedia and Expo (ICME), IEEE International Conference on, pp.290 ậ 295 [βγ] Gallas, A., Barhoumi, W., Zagrouba, on Wavelet E (β01β),‘Image Retrieval Based Sub-bands and Fuzzy Weighted Regions’, Communications and Information Technology (ICCIT), 2012 International Conference on , pp: 33 ậ 37 [24] Quellec, G., Lamard, M., Cochener, B., Roux, C., Cazuguel, G.(2012), ‘Comprehensive Wavelet-Based Image Characterization for Content Based Image Retrieval’, Content-Based Multimedia Indexing (CBMI), 2012 10th International Workshop on , pp:1-6 [β5] Anil Balaji Gonde, R.P Maheshwari and Balasubramanian (β01γ),‘Modified curvelet transform with vocabulary tree for content based image retrieval’, Digital Signal Processing vol 23, pp: 142ậ150 53 [26] Esmat Rashedi, Hossein Nezamabadi-pour and Saeid Saryazdi (β01γ),‘A simultaneous feature adaptation and feature selection method for content-based image retrieval systems’, Knowledge-Based SystemsVolume 39, Pages 85ậ94 [27] Ela Yildizer, Ali (β01β),‘Efficient Metin Balci, Mohammad content-based image Hassan and Reda retrieval using Alhajj Multiple Support Vector Machines Ensemble’, Expert Systems with ApplicationsVolume γ9, Issue 3, Pages 2385ậ2396 [β8] Quellec, G., Lamard, M., Cazuguel, G., Cochener, B (β01β), ‘Fast WaveletBased Image Characterization for Highly Adaptive Image Retrieval’, Ieee Transactions On Image Processing, Vol 21, No [β9]Tamura, H., Mori, S., and Yamawaki, T (1978), ‘Textural Features Corresponding to Visual Perception’, IEEE Trans Systems, Man Cybernetics, vol 8, pp 460ậ472 [γ0] Howarth, P and Rüger, S (β005), ‘Robust Texture Features for Still Image Retrieval’, IEEE Proc Vision, Image Signal Processing, vol 15β, no 6, pp 868ậ 874 [γ1] Howarth, P and Rüger, S (β004), ‘Evaluation of Texture Features for Contentbased Image Retrieval’, Proc Of CIVR'04, pp γβ6ậ334 [γβ] Sebe, N and Lew, M.S.(β000), ‘Wavelet Based Texture Classification, Proc of Int Conf on Pattern Recognition’, vol γ, pp 959ậ962 [γγ] Manjunath, B.S and Ma, W.Y (1996), ‘Texture Features for Browsing and Retrieval of Image Data’, IEEE Trans Pattern Analysis Machine Intelligence, vol 18, no 8, pp 837ậ842 [γ4] Manjunath, B.S., Wu, P., Newsam, S., and Shin, H.D (β000), ‘A Texture Descriptor for Browsing and Similarity Retrieval, Proc Signal Processing Image Commun.’, nos 1ậ2, pp 33ậ43 [γ5] Bell, A.J and Sejnowsky, T.J (1997), ‘The ‘Independent Components’ of Natural Scenes are Edge Filters, Vision Research’ , no γ7, pp γγβ7ậ3338 54 [36] Borgne, H., Guerin-Dugue, A., and Antoniadis, A.(β004), ‘Representation of Images for Classification with Independent Features, Pattern Recognition Letters’, vol β5, pp 141ậ 154 [γ7] Snitkowska, E and Kasprzak, W (β006), ‘Independent Component Analysis of Textures in Angiography Images, Computational Imaging Vision’, vol γβ, pp 367ậ 372 [38] Field, D.J (1987), Relations Between the Statistics of Natural Images and the Response Properties of Cortical Cells,J Optical Soc America, vol 12, no 4, pp 2370ậ 2393 [39] Chuen-Horng Lin, Rong-Tai Chen, Yung-Kuan Chan (β009), ‘A smart contentbased image retrieval system based on colour and texture feature’, Image and Vision Computing, vol 27, 658ậ665 [40] Jun Yue, Zhenbo Li, Lu Liu and Zetian Fu(β011), ‘Content-based image retrieval using colour and texture fused features’, Mathematical and Computer Modelling, vol 54, pp 1121ậ 1127 [41] Xiang-Yang Wang & Bei-Bei Zhang & Hong-Ying Yang, ‘Content-based image retrieval by integrating colour and texture features’, Multimed Tools Appl, DOI 10.1007/s11042-012-1055-7 [4β] H Abrishami Moghaddam and M Nikzad Dehaji, ‘ Enhanced Gabor wavelet correlogram feature for image indexing and retrieval’, Pattern Anal Applic [4γ] Ela Yildizer, Ali Metin Balci, Tamer N Jarada, Reda Alhajj, ‘Integrating wavelets with clustering and indexing for effective content-based image retrieval’, Knowledge-Based Systems 31 (2012) 55ậ66 [44] Dengsheng Zhang - M Monirul Islam - Guojun Lu and Ishrat Jahan Sumana, ‘ Rotation Invariant Curvelet Features for Region Based Image Retrieval’, Int J Comput Vis (2012) 98:187ậ201 [45] Ela Yildizer, Ali Metin Balci, Mohammad Hassan, Reda Alhajj, ‘Efficient content-based image retrieval using Multiple Support Vector Machines Ensemble’, Expert Systems with Applications γ9 (2012) 2385ậ239 55 [46] G Quellec, ‘Wavelet M Lamard, optimization for G Cazuguel, content-based B Cochener, image retrieval C in Roux, medical databases’, Medical Image Analysis 14 (β010) ββ7ậ241 [47] Liang-Hua Chen, Yao-Ling Hung, and Li-Yun Wang (β01β), ‘An Integrated Approach to Image Retrieval’, Telecommunications and Signal Processing (TSP), 2012 35th International Conference on, pp: 695 ậ 699, 2012 [48] Meng Fanjie, Guo Image Retrieval Baolong and Wu Xianxiang (β01β), ‘Localized Based on Interest Points’, β01β International Workshop on Information and Electronics Engineering (IWIEE), pp 3371 ậ 3375 [49] H Abrishami Moghaddam and M Nikzad Dehaji (β01γ), ‘Enhanced Gabor wavelet correlogram feature for image indexing and retrieval’, Pattern Analysis and Applications, Vol 16, issue 2, pp:163-177 [50] Teague, M (1980), ‘Image Analysis via the General Theory of Moments’, J Optical Society America, vol 70, no 8, pp 920ậ930 [51] Hu, M K (196β), ‘Visual Pattern Recognition by Moment Invariants’, IEEE Trans Information Theory, vol 8, issue 2, pp 179ậ187 [5β] Luren, Y and Fritz, A (1994), ‘Fast Computation of Invariant Geometric Moments: A New Method Giving Correct Results’, Proc of IEEE Int Conf on Image Processing [5γ] Hew, P., Geometric and Zernike Moments (1996), ‘Diary’, Department of Mathematics, The University of Western Australia, 1996 http://citeseer.ist.psu.edu/hew96 geometric.html [54] Zhang, D.S and Lu, G., ‘Generic Fourier Descriptor for Shape-based Image Retrieval (β00β)’, Proc of IEEE Int Conf on Multimedia and Expo (ICME2002), Lausanne, Switzerland, vol 1, pp 425ậ428 [55] R Krishnamoorthy, S Sathiya Devi 9β01γ), ‘Image retrieval using edge based shape similarity with multiresolution enhanced orthogonal polynomials model', Digital Signal Processing vol 23, 555ậ568 56 [56] Z M Ma, Gang Zhang and Li Yan(β011), ‘Shape feature descriptor using modified Zernike moments’, Pattern Anal Applic vol 14, pp 9ậ22 [57] Anjali Goyal, Ekta Walia, ‘Variants of dense descriptors and Zernike moments as features for accurate shape-based image retrieval’, SIViP DOI 10.1007/s11760-012-0353-x [58] Meng Fanjie, Guo Baolong, Wu Xianxiang (β01β), ‘Localized Image Retrieval Based on Interest Points’, Procedia Engineering, vol β9 pp γγ71ậ3375 [59] Flusser J (2005) Moment invariants in image analysis Enformatika 11 [60] Kotoulas L, Andreadis I (2005) Image analysis using moments 5th International Conference on Technology and Automation, Thessaloniki, Greece 360ậ364 ... nghiên c u phát tri n h th ng tìm ki m nh ngày tr nên c p thi t Có hai ki u tìm ki m lƠ tìm ki m theo t khóa tìm ki m theo n i dung nh (CBIR-Content Based Image Retrieval), tìm ki m theo t khóa d th... cho nhu c u c a ng i M c dù m t s mơ hình tìm ki m nh đư đ i đáp ng ph n nhu c u tìm ki m nh, song nâng cao ch t l ng tìm ki m v n đ đ c đ t Bài tốn tìm ki m nh, nâng cao ch t l ng x p h ng nh... tr ng pháp nh m tìm c t p d li u nh l n 1.4 Gi i h n c aăđ tài Nghiên c u m t s k thu t tìm ki m nh, đ c bi t lƠ k thu t tìm ki m nh d a vƠo mƠu s c, c u trúc vƠ n i dung c a nh Tìm ki m nh d a