(3.237.97.64) 您好!臺灣時間:2021/03/04 10:43
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:陳凱宏
研究生(外文):Kai-Hung Chen
論文名稱:以顏色、紋理及空間關係為基礎之影像查詢系統
論文名稱(外文):A Color-Space and Color-Texture Based Image Retrieval System
指導教授:林春宏林春宏引用關係詹永寬詹永寬引用關係
指導教授(外文):Chuen-Horng LinYung-Kuan Chan
學位類別:碩士
校院名稱:國立臺中技術學院
系所名稱:資訊科技與應用研究所
學門:電算機學門
學類:電算機應用學類
論文種類:學術論文
論文出版年:2006
畢業學年度:95
語文別:英文
論文頁數:62
中文關鍵詞:影像查詢顏色紋理K-meansQRCBIR
外文關鍵詞:content-based image retrievalPCAcolor-based image retrievaltexture-based image retrievalspatial-based image retrieval
相關次數:
  • 被引用被引用:0
  • 點閱點閱:154
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
本論文提出兩種萃取影像(image)特徵(feature)的方法來做為影像查詢的依據。此外,也提出一套相似影像過濾(filter)技術,作為影像比對時的初步篩選,藉以縮短系統在查詢影像時所需之比對時間。第一個影像特徵為影像顏色(color)及紋理(texture)特徵,本文稱為多向顏色複雜度(multi-orientation and resolution color complexity, MORCC)。第二個影像特徵是顏色空間(color space)分布特徵的取出技術,稱之為顏色空間關係(color spatial relation, CSR)。
MORCC係先將一個影像相鄰像素(pixel)值差值的絕對值取出,其中相鄰像素可以是水平、垂直與斜角等方向的相臨1個、2個、…、n個像素。為判別不同差值大小所代表影像顏色及紋理的意義,MORCC將不同差值依照大小區分成數個群組(group)。最後,再以統計方法對每個群組進行分析,以此做為此影像顏色及紋理的特徵值。CSR則使用K-means分群演算法將影像所有的像素值分成數個群組,接著再計算各群組內所有像素之間的空間位置距離的總和,做為此影像的顏色空間分布的特徵。此外,為更進一步表現影像的紋理特性及顏色空間關係,本論文也嘗試將MORCC及CSR特徵作進一步修正,額外提出了IMORCC及FCSR特徵。此兩種特徵皆是MORCC及CSR特徵的細部修正,更能描述影像紋理及顏色分佈的特性。
影像的特性與內容不同,代表著此影像擁有不同的特徵,有些影像對顏色及紋理特徵比較明顯,而有些影像對顏色空間特徵比較敏感。因此,本文結合MORCC及CSR的特徵來做為影像查詢。此外,本論文也提出一套權重(weight)值自動產生器,分別訓練出最適合MORCC及CSR特徵之權重,更有效的改善查詢的正確率。
為加速查詢影像與影像資料庫的特徵值之比對,本論文也提出了一個相似影像過濾技術。針對影像資料庫內相似的影像做初步的篩選,保留少部分比較相似的影像,作為本查詢系統最後詳細部分的比對。主要的技術係將影像資料庫所有影像之CSR的特徵,依K-means演算法分成數個群組,而查詢影像CSR的特徵值將分別與這些群組做歐式距離計算,並選擇距離最小的數個群組,做為本文相似影像過濾的依據。最後,本文實驗的部分將依MORCC特徵值、CSR特徵值、結合MORCC及CSR特徵值、IMORCC特徵值、FCSR特徵值、結合IMORCC及FCSR特徵和套用權重值自動產生器及過濾器做一系列的比較與分析。
This thesis proposes two image features, multi-orientation and resolution color complexity (MORCC) and color-space relation (CSR). MORCC computes the differences of pixel colors in multiple orientations and resolutions, which can describe the variation of textures of an image in different colors. CSR depicts the spatial distribution of similar color pixels in an image. Firstly, the variation of pixel colors was separated into several groups. The mean, standard and skewness of the pixels in each group were computed as MORCC feature. Secondly, the pixels of the image were separated into several clusters according to their color similarity, and then the average coordinates of the pixels in each cluster were computed. It calculates the summation of the distances between the average coordinates and each pixel in the same cluster. This summation can describe the spatial distribution of the pixels in the cluster. Owing to a high complimentarily between these two features, this thesis integrates both features to develop a color-space and color-texture based image retrieval system (CSCTIR system). The CSCTIR system can recognize both images with similar colors, textures, and spatial distributions.
In order to further describing the texture attributes and color space relation, two modified approaches were also proposed: improved MORCC (IMORCC) and fast CSR (FCSR). Both features were originated from the MORCC and CSR features.
Considering there are many features which make system performance poor, a feature selector based on principle component analysis (PCA) for eliminating undesired features were also proposed. Besides, to well combine MORCC and CSR features, an automatic weight generator are presented. The weight generator can generate the most suitable values for MORCC and CSR features automatically in a very short time.
Finally, in order to make the time during similarity comparison shorter, we also proposed a filter based on CSR features for its fewer feature vectors than MORCC. Firstly, CSR features of database images are clustered into several groups. By calculating the distance between CSR features of query images and these groups, we can rule out some groups with larger distance values. Images in the rest groups are those we want for final comparison. Through many restrict experiments with 4 image sets; the results demonstrate that the integration of these features really makes our retrieval system excellent accuracy and performance. In addition, PCA-based feature selector scheme also rule out most useless features which makes retrieval time much faster than that with full image features. Finally, the proposed filter also plays good role in the elimination of dissimilar database image.
Abstract in Chinese
Abstract in English
Acknowledgement in Chinese
List of Tables
List of Figures

Chapter 1 Introduction
1.1 Background
1.2 Motivation
1.3 Framework and Workflow
1.4 Thesis Organization

Chapter 2 Related Works
2.1 Color Histogram
2.2 Image Retrieval by Texture Similarity
2.3 Content-Based Image Retrieval Using Growing Hierarchical Self-
Organizing Quadtree Map
2.4 Wavelet Correlogram: A New Approach for Image Indexing and
Retrieval
2.5 Content Based Image Retrieval Using Motif Cooccurrence Matrix

Chapter 3 Color-Space and Color-Texture Based Image Retrieval System
3.1 MORCC Feature Extraction
3.2 CSR Feature Extraction
3.3 CSCTIR System
3.4 Experimental Results and Comparisons
3.4.1 The Performance of MORCC Feature
3.4.2 The Performance of CSR Feature
3.4.3 The Performance of Combining MORCC and CSR Features

Chapter 4 Improved Color-Space and Color-Texture Based Image Retrieval
System
4.1 Improved MORCC (IMORCC) Feature Extraction
4.2 Fast CSR (FCSR) Feature Extraction
4.3 Feature Selector
4.4 Auto-Weight Generator
4.5 ICSCTIR System
4.6 Experimental Results and Comparisons
4.6.1 The Performance of IMORCC Feature
4.6.2 The Performance of FCSR Feature
4.6.3 The Performance of Feature Selector
4.6.4 The Performance of Combining IMORCC and FCSR Features
4.6.5 Comparing ICSCTIR System with Hung and Dai’s Approach
4.6.6 Robustness of ICSCTIR Image Retrieval System
4.7 The Performance of ICSCTIR System on Other Image Sets
4.7.1 The Performance of ICSCTIR System on Image Set 2
4.7.2 The Performance of ICSCTIR System on Image Set 3
4.7.3 The Performance of ICSCTIR System on Image Set 4
4.7.4 The Performance of ICSCTIR System Comparing to
Wavelet Correlogram
4.7.5 Robustness of ICSCTIR System on Image Set 4

Chapter 5 Clustering-Based Filter
5.1 CSR Based Filter
5.2 Fast CSR Based Filter
5.3 Experimental Results of the Filters
5.3.1 The Performance of Filter on CSCTIR System
5.3.2 The Performance of Fast Filter on ICSCTIR System

Chapter 6 Conclusions
[1]Y. Gong, H. Zhang, H. Chuant, and M. Skauuchi “An Image Database System with Contents Capturing and Fast Image Indexing Abilities,” Proceedings of IEEE International Conferences on Multimedia Computing and Systems, Boston, Massachusetts, USA, pp. 121–130, May, 1994.

[2]V. N. Gudivada, and V. V. Raghavan, “Design and Evaluation of Algorithms for Image Retrieval by Spatial Similarity,” ACM Transactions on Information Systems (TOIS), Vol. 13, No. 2, pp. 115-144, 1995.

[3]Y. Gu, B. Panda, and K. A. Haque, “Design and Analysis of Data Structures for Querying Image Databases,” ACM Symposium on Applied Computing, Las Vegas, Nevada, USA, pp. 236-241, 2001.

[4]Y. Rui, and T. S. Huang, “Image Retrieval: Current Techniques, Promising Directions, and Open Issues,” Journal of Visual Communication and Image Representation, Vol. 10, pp. 39–62, October, 1999.

[5]M. Stricker, and M. Orengo, “Similarity of color images,” SPIE: Storage Retrieval Image and Video Database III, Vol. 2420, pp. 381–392, February, 1995.

[6]P. W. Huang, and S. K. Dai, “Image Retrieval by Texture Similarity,” Pattern Recognition, Vol. 36, No. 3, pp. 665-679, 2003.

[7]SitaoWu, M.K.M. Rahman and Tommy W.S. Chow, “Content-based image retrieval using growing hierarchical self-organizing quadtree map,” Pattern Recognition, Vol. 38, pp. 707-722, 2005.

[8]H. Abrishami Moghaddam, T. Taghizadeh Khajoie, A.H. Rouhi, and M. Saadatmand Tarzjan, “Wavelet correlogram: A new approach for image indexing and retrieval,” Pattern Recognition, Vol. 38, pp. 2506-2518, 2005.

[9]R. Brnuelli, and O. Mich, “Histograms Analysis for Image Retrieval,” Pattern Recognition, Vol. 34, pp. 1625-1637, 2001.

[10]J. M. Fuertes, M. Lucena, N. Peres de la Blanca and J. Chamorro-Martinez, “A Scheme of Color Image Retrieval From Databases,” Pattern Recognition, Vol. 22, No. 3, pp. 323-337, 2001.

[11]Y. D. Chun, S. Y. Seo, and N. C. Kim, “Image Retrieval Using BDIP and BVLC Moments,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 13, Issue 9, pp. 951-957, 2003.

[12]B. C. Ko, and H. Byun, “FRIP: A Region-Based Image Retrieval Tool Using Automatic Image Segmentation and Stepwise Boolean AND Matching,” IEEE Transactions on multimedia, Vol. 7, No. 1, pp. 105-113, 2005.

[13]Y. K. Chan and C. Y. Chen, “Image Retrieval System Based on Color-Complexity and Color-Spatial Features,” The Journal of Systems and Software, Vol. 71, Issue 1-2, pp. 65-70, 2004.

[14]C. C. Chang and Y. K. Chan, “A Fast Filter for Image Retrieval Based on Color Features,” SEMS2000, Baden-Baden, German, pp. 47-51, 2000.

[15]H. Nezamabadi-Pour and E. Kabir, “Image Retrieval Using Histograms of Uni-color and Bi-color Blocks and Directional Changes in Intensity Gradient”, Pattern Recognition Letters, Vol. 25, Issue 14, pp.1547-1557, 2004.

[16]S. Liapis, and G. Tziritas “Color and Texture Image Retrieval Using Chromaticity Histograms and Wavelet Frames,” IEEE Transactions on Multimedia, Vol. 6, No. 5, pp. 676-686, 2004.

[17]R. M. Haralick, B. Shanmugam, and I. Dinstein, “Texture Features for Image Classification,” IEEE Transactions on Systems, Man, and Cybernetics, Vol. 3, No. 6, pp. 610–621, November, 1973.

[18]N. Jhanwar, S. Chaudhurib, G. Seetharamanc, and B. Zavidovique, “Content Based Image Retrieval Using Motif Co-occurrence Matrix,” Image and Vision Computing, Vol. 22, pp. 1211–1220, 2004.

[19]B. M. Mehtre, M. Kankanhalli, and W. F. Lee, “Shape Measures for Content-Based Image Retrieval: A Comparison Info,” Processing & Management, Vol. 33, pp. 319–337, 1997.

[20]M. J. Swain, and D. H. Ballard, “Color Indexing,” International Journal of Computer Vision, Vol. 7, pp. 11-32, 1991.

[21]I. Daubechies, “Orthonormal Bases of Compactly Supported Wavelets,” Communications on Pure and Applied Mathematics, Vol. 41, pp. 909-996, 1998.

[22]R. M. Haralick, and L. G. Shapiro, “Computer and Robot Vision,” Vol. I, Addison-Wesley, Reading, MA, 1992.

[23]M. C. Su, and C. H. Chou, “A Modified Version of the K-means Algorithm with a Distance Based on Cluster Symmetry,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, No. 6, pp. 674-680, 2001.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔