進階搜尋


 
系統識別號 U0026-0812200910414424
論文名稱(中文) 統計分水嶺影像分割之研究
論文名稱(英文) A Study on Statistical Watershed Segmentation
校院名稱 成功大學
系所名稱(中) 統計學系碩博士班
系所名稱(英) Department of Statistics
學年度 91
學期 2
出版年 92
研究生(中文) 王怡穎
研究生(英文) Yi-Ying Wang
電子信箱 yyw.ncku@msa.hinet.net
學號 r2690406
學位類別 碩士
語文別 英文
論文頁數 68頁
口試委員 指導教授-孫永年
指導教授-吳鐵肩
口試委員-李隆安
中文關鍵字 分水嶺分割  馬可夫模型  多維解析度  合併 
英文關鍵字 Multi-resolution  Markov random field  Merge  Watershed segmentation 
學科別分類
中文摘要 在這篇文章中,我們研究的主題集中在影像分析上。影像分割是影像分析的一個重要的過程。我們結合了分水嶺分割法和統計方法找出所感興趣的區域。用分水嶺分割法,有一個缺點就是所分割出來的區域過多,所以我們的目標是用統計方法把過多的區域合併,以達到更好的視覺效果。
首先,為了去掉雜訊,我們用中位數過濾法,使原始影像更平滑。然後由這張影像算出梯度影像,接著用這張梯度影像當成要進行分水嶺分割的影像。正如所預期的,用分水嶺分割出來的影像呈現一區區的小區域,於是我們就用統計方法來解決這個問題。
不同於傳統的方法,我們應用統計方法來處理合併的問題。在合併的第一個步驟,利用多維解析度的概念,找出有信心的邊,作為合併前的前處理步驟。之後,用三種標準逐一地判斷是否要合併小區域,這些標準是經由計算區域的特性而來的。其中,我們有考慮區域間平均數的差異、區域間的變異是否相同、用馬可夫模型來檢測邊有沒有存在的必要。
在這篇文章中,背景介紹在第一章;分水嶺分割法的介紹在2.1節;多維解析度的概念在2.2節;相關的統計方法在第三章;第四章介紹整個合併的流程;實驗結果在第五章;一些討論和結論在第六章。
英文摘要 In this thesis, the area of research is focused on computer image analysis. Image segmentation is one of the most important processes in computer image analysis. In this research, we proposed a hybrid image segmentation method consisting of the watershed technique and a statistic merging process to detect areas of interest from a given image. As this initial segmentation by watershed approach usually suffers from an over-segmented image, our aim is to design the statistic merging mechanism and improve the segmentation results.
At first, in order to reduce the noise, we apply the median filter on the input image. Then, we acquire the gradient image from the resulting image as the new input image, which is then used to implement the watershed transform. The watershed transform obtains an over-segmented result as expected. Therefore, we have to utilize the proposed merge procedure.
Different from the conventional methods, we want to use the statistical method to deal with the merge problem. The merge method employs a multi-resolution procedure together with three merging criteria. A simple multi-resolution edge detection technique is designed as the preprocessing step. Three merging criteria based on the region properties are defined. These criteria are calculated for every small region of the watershed result. Then, we can check the similarity of the adjacent regions between every pair of neighboring regions. By using the Markov random field (MRF) model, we obtain the optimal segmentation by merging regions based on the selected criteria.
In this thesis, background introduction is given in Chapter 1. The watershed transform is described in Section 2.1. The concept of multi-resolution is depicted in Section 2.2. The related statistical methods are presented in Chapter 3. Chapter 4 illustrates the merging procedure. Experimental studies are demonstrated in Chapter 5 with five standard images. And some discussions and conclusions are given in Chapter 6.
論文目次 Chapter 1 Introduction………………..……………….…….1
1.1 Introduction……………………………………...…..1
1.2 Threshold techniques…….……………………..…...3
1.3 Edge-based methods………….…….…...…..……....5
1.4 Region-based methods……….………..………….…6
1.5 Hybrid-methods…………..………....………….…...7
Chapter 2 Watershed algorithm………………………….…9
2.1 Watershed Algorithm…………………………….….9
2.2 Multi-Resolution…………………………..………...…..18
Chapter 3 Statistical methods………….………………..…23
3.1 Bayesian approach………………..…………..……23
3.2 Bartlett test……………….….……………….…….21
3.3 Markov random field………….……………..…….25
3.4 Normal distribution transform………………....…..32
Chapter 4 Merging procedure………………………….…….44
Chapter 5 Experiment results…………………………….…..51
Chapter 6 Discussions and Conclusions………………….….63
Reference………………………………………………………64
Appendix…….…………………………………………...…….67
參考文獻 [1] R. Haralick and L. Shapiro, ‘Image segmentation techniques,’ CVGIP, Vol. 29, pp.100-132 (1985).
[2] K. Mardia and T. Hainsworth, ‘A spatial thresholding method for image
segmentation,’ IEEE Trans. Pattern Anal. Machine Intell., Vol. 10, pp.
919-927, Nov. (1988).
[3] http://sern.ucalgary.ca/courses/cpsc/533/w02/Perception/Perception.ppt
[4] A. Jain, Fundamentals of Digital Image Processing. Englewood Cliffs, NJ; Prentice-Hall, (1989).
[5] V. Nalwa, A Guided Tour of Computer Vision. Reading, MA: Addison-
Wesley, (1993).
[6] J. Canny, ‘A computational approach to edge detection,’ IEEE Trans.
Pattern Anal. Machine. Intell., Vol. PAMI-8, pp. 679-698, Nov. (1986).
[7] D. Marr and E. Hildreth, ‘Theory of edge detection,’ in Proc. R. Soc. Lond. B, no. 207, pp. 187-217. (1980).
[8] S. Horowitz and T. Pavlidis, “Picture segmentation by a tree traversal
algorithm,” J. Assoc. Comput. Mach., Vol. 23, pp. 368-388, Apr. (1976).
[9] S. Chen, W. Lin, and C. Chen, ‘Split-and-merge image segmentation
based on localized feature analysis and statistical tests,’ CVGIP: Graph.
Models Image Process., Vol. 53, pp. 457-475, Sept. (1991).
[10] P. Besl and R. Jain, ‘Segmentation through variable-order surface
fitting,’ IEEE Trans. Pattern Anal. Machine Intell., Vol. 10, pp. 167-192,
Mar. (1988).
[11] T. Pavlidis and Y. Liow, ‘Integrating region growing and edge detection,’
IEEE Trans. Pattern Anal. Machine Intell., Vol. 12, pp. 225-233,
Mar. (1990).
[12] L. D. Griffin, A. C. F. Colchester, and G. P. Robinson, ‘Scale and
segmentation of grey-level images using maximum gradient paths,’
Image Vis. Comput., Vol. 10, pp. 389-402, July/Aug. (1992).

[13] F. Meyer and S. Beucher, ‘Morphological segmentation,’ J. Vis. Commun.
Image Represent., Vol. 1, pp. 21-46, Sept. (1990).
[14] J. M. Gauch and S. M. Pizer, ‘Multiresolution analysis of ridges and
valleys in gray-scale images,’ IEEE Trans. Pattern Anal. Machine Intell.,
Vol. 15, pp. 635-646, June (1993).
[15] Kostas Haris, Serafim N. Efstratiadis, Member, IEEE, Nicos Maglaveras, Member, IEEE,and Aggelos K. Katsaggelos, Fellow, IEEE, ‘Hybrid Image Segmentation Using Watersheds and Fast Region Merging,’ IEEE Transactions On Image Processing, Vol. 7, no. 12, December (1998).
[16] L. Vincent and P. Soille, ‘Watersheds in digital spaces: An efficient
algorithm based on immersion simulations,’ IEEE Trans. Pattern Anal.
Machine Intell., Vol. 13, pp. 583-598, June (1991).
[17] C.-T. Li, ‘Multiresolution image segmentation integrating Gibbs sampler
and region merging algorithm,’ Signal Processing, 83, pp.67-78 (2003).
[18] Jong-Bae Kim, Hang-Joon Kim, ‘Multiresolution-based watersheds for efficient
image segmentation,’ Pattern Recognition Letters, 24, pp.473-488 (2003).
[19] Richard R. Schultz, Student Member, IEEE, and Robert L. Stevenson, Member, IEEE, ‘A Bayesian Approach to Image Expansion for Improved Definition,’ IEEE Transactions On Image Processing, Vol. 3, no. 3, May (1994).
[20] James O. Berger. Statistical Decision Theory and Bayesian Analysis (1985).
[21] Bartlett, M. S., Properties of sufficiency and statistical tests. Proceedings of the Royal Statistical Society Series A 160, pp. 268–282 (1937).
[22] Stan Z. Li. Markov Random Field Modeling in Image Analysis. Computer Science Workbench Series Editor: Tosiyasu L. Kunii (2001).
[23] Geman, S. and Geman, D. ‘Stochastic relaxation, Gibbs distribution and the Bayesian restoration of images,’ IEEE Transactions on Pattern Analysis and Machine Intelligence, 6(6): pp. 721-741 (1984).
[24] Derin, H. and Elliott, H. ‘Modeling and segmentation of noisy and textured images using Gibbs random fields,’ IEEE Transactions on Pattern Analysis and Machine Intelligence, 9(1): pp. 39-55 (1987).
[25] Geiger, D. and Girosi, F, ‘Parallel and deterministic algorithms from MRF's: surface reconstruction,’ IEEE Transactions on Pattern Analysis and Machine Intelligence, 13(5): pp.401-412. (1991)
[26] Rangarajan, A. and Chellappa, R., ‘Generalized graduated non-convexity algorithm for maximum a posteriori image estimation,’ In Proceedings of International Conference Pattern Recognition, pp. 127-133 (1990).
[27] Shapiro, S. S. and Wilk, M. B., ‘An analysis of variance test for normality (complete samples)’, Biometrika, 52, 3 and 4, pp. 591-611 (1965).
[28] Box, G.E.P. and D.R. Cox, An analysis of transformations (with discussion). J. Royal Statist. Soc. Ser. B, 26: pp. 211-246 (1964).
[29] Anjan Sarkar, Manoj K. Biswas, and K. M. S. Sharma, ‘A Simple Unsupervised MRF Model Based Image Segmentation Approach’ IEEE Transactions On Image Processing, Vol. 9, no. 5, May (2000).
[30] Michael W. Hansen and William E. Higgins, ‘Watershed-Based Maximum-Homogeneity Filtering,’ IEEE Transactions On Image Processing, Vol. 8, no. 7, July (1999).
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2004-07-25起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2004-07-25起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw