臺灣博碩士論文加值系統

在大量的連續影像中分割出3D物體需要龐大的計算量（記憶體，計算時間等）。
此外，對操作影像處理的研究員來講，
在不同的案例間調整參數是一件重複且冗長的工作。
有些時候甚至會因為時間消耗的問題導致整個處理過程幾乎不可能在預想的時間點完成。
舉例來說，我們都知道重建三維的數位影像應用是一件有挑戰性的工作。
這件複合的任務可以被分成數個不同的影像處理主題。並且每個影像處理都有自己的參數。
理論上，越多處理應該要得到更準確的結果，但事實上，整個研究的過程有可能因此走到瓶頸。
因為影響結果的因素實在太多導致分析出來問題在哪變得困難重重。在本篇論文中，
我們發展了一個創新的方法去探測和從原始資料中將影像物體分割出來，其中我們提出並實作「哨兵」的觀念及其應用。
經由以哨兵為基礎的影像分割結果，較以往的研究結果更為可靠且需要較少的人為參數調整。
最後，我們設計了一個視覺化的工具，以物件導向為基礎輔助使用者以較直覺的方式完成3D影像的工作。

Segmenting a huge three-dimensional(3D) image or a sequence of two-dimensional(2D)images requires tremendous computation.
Also, it is tedious for the imaging operator to adjust parameters for different study cases.
Sometimes the time-consuming issue also made the whole procedure nearly impossible to accomplish on time.
For example, we all know that reconstructing (3D) digital image application is a challenging job.
This complex task can be divided into different topics of image processing.
And each processing step has its own parameters.
Ideally, we expect more precise and meaningful results with more processing steps.
But in fact, the research procedure may run into the bottleneck just because of the factors are too many to handle, and can not be easily analyzed.
In this thesis, we developed an innovative method to probe and segment object from raw image data.
Besides, we developed an image visualization platform for researching segmentation method of sequential images.
And with the proposed software tools, we can visualize the data to screen.
The significant feature and main contribution of this thesis is the sentinel concept.
With it, we can have two criteria in the same time and let the segment processing more adapted to tough condition.

Acknowledgments IV
Abstract (Traditional Chinese) V
Abstract VI
1 Introduction 1
1.1 An Introduce of Our Experiment Procedure . . . . . . . 3
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Literature Review 7
2.1 Background . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 The capability estimation . . . . . . . . . . . . . 7
2.2 Processing Model . . . . . . . . . . . . . . . . . . . . . 8
2.3 De-noise or smooth . . . . . . . . . . . . . . . . . . . . 8
2.4 Image histogram and image histogram equalization . . . 9
2.5 Segmentation . . . . . . . . . . . . . . . . . . . . . . . 11
2.5.1 K-mean clustering . . . . . . . . . . . . . . . . 11
2.5.2 Histogram based methods . . . . . . . . . . . . 12
2.5.3 Region-based methods . . . . . . . . . . . . . . 12
3 A Self-Constrained Image Segmentation Mechanism 14
3.1 Problem Statement . . . . . . . . . . . . . . . . . . . . 15
3.2 Growing, Connecting . . . . . . . . . . . . . . . . . . . 16
3.3 Memory Using Issue : Stack overflow . . . . . . . . . . 16
3.4 Probe . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.4.1 3D Probing . . . . . . . . . . . . . . . . . . . . 19
3.5 Pushing pitfall . . . . . . . . . . . . . . . . . . . . . . . 26
VII
3.6 Sentinel . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.6.1 Statistical discuss . . . . . . . . . . . . . . . . . 28
3.6.2 Orthogonal voting . . . . . . . . . . . . . . . . 30
3.6.3 Abstract discussion . . . . . . . . . . . . . . . . 31
4 Math model 34
4.1 Foundation model . . . . . . . . . . . . . . . . . . . . . 34
4.2 Bridge-wall dilemma . . . . . . . . . . . . . . . . . . . 37
5 Visualization tools 39
5.1 Environment . . . . . . . . . . . . . . . . . . . . . . . 39
5.2 3D visualization pipeline . . . . . . . . . . . . . . . . . 46
6 Sentinel Experimental Result 47
6.1 Mouse liver . . . . . . . . . . . . . . . . . . . . . . . . 47
6.2 Heart and blood vessel . . . . . . . . . . . . . . . . . . 50
7 Conclusion and Future work 53
7.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 53
7.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . 53
References 54

List of Figures
1.1 The MRI equipment and CT equipment . . . . . . . . . 2
1.2 The system model . . . . . . . . . . . . . . . . . . . . . 3
1.3 Decomposing procedure . . . . . . . . . . . . . . . . . 4
1.4 Slicer . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 The Boundary is ambiguous and blurry . . . . . . . . . . 5
1.6 The General Processing . . . . . . . . . . . . . . . . . . 6
1.7 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 Region-Growing . . . . . . . . . . . . . . . . . . . . . 13
3.1 Sobel masks . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2 3-D sobel, relationship map . . . . . . . . . . . . . . . . 21
3.3 X-y coordinate vectors, 3x3 relationship map . . . . . . 22
3.4 Trigger Bar . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5 Chebyshev's inequality . . . . . . . . . . . . . . . . . . 29
3.6 Empirical rule . . . . . . . . . . . . . . . . . . . . . . . 30
3.7 Orthogonal probing . . . . . . . . . . . . . . . . . . . . 31
3.8 Circle growing . . . . . . . . . . . . . . . . . . . . . . 32
3.9 Two criteria . . . . . . . . . . . . . . . . . . . . . . . . 32
3.10 Decision Flowchart . . . . . . . . . . . . . . . . . . . . 33
3.11 Voting behavior will stop the growing . . . . . . . . . . 33
4.1 The connecting evidence between pixels . . . . . . . . . 38
5.1 An overall flowchart . . . . . . . . . . . . . . . . . . . 39
5.2 Our software outlooking . . . . . . . . . . . . . . . . . 40
5.3 Empty volume . . . . . . . . . . . . . . . . . . . . . . . 40
5.4 Filled volume . . . . . . . . . . . . . . . . . . . . . . . 40
5.5 Function with label . . . . . . . . . . . . . . . . . . . . 41
5.6 Histogram . . . . . . . . . . . . . . . . . . . . . . . . . 42
IX
5.7 Sentinel segmentation tab . . . . . . . . . . . . . . . . . 43
5.8 Sentinel control panel . . . . . . . . . . . . . . . . . . . 44
5.9 3D visualization . . . . . . . . . . . . . . . . . . . . . . 45
6.1 Original images . . . . . . . . . . . . . . . . . . . . . . 47
6.2 Probing using sobel masks,tolerance 20,send sentinel when
100% tolerance . . . . . . . . . . . . . . . . . . . . . . 48
6.3 Probing using sobel masks, tolerance 20,send sentinel
when 70% tolerance . . . . . . . . . . . . . . . . . . . 49
6.4 Probing using sobel masks,tolerance 20,send sentinel when
60% tolerance . . . . . . . . . . . . . . . . . . . . . . . 49
6.5 Probing using sobel masks,tolerance 20,send sentinel when
50% tolerance . . . . . . . . . . . . . . . . . . . . . . . 50
6.6 Probing using sobel masks,tolerance 20,send sentinel when
90% tolerance . . . . . . . . . . . . . . . . . . . . . . . 50
6.7 Probing using sobel masks,tolerance 20,send sentinel when
90% tolerance . . . . . . . . . . . . . . . . . . . . . . . 51
6.8 The blood vessel growing is deep . . . . . . . . . . . . . 52
6.9 The blood vessel growing is restricted . . . . . . . . . . 52

References
[1] Acharya and Ray. ``Image Processing: Principles and Applica-
tions''. Wiley-Interscience, 2005.
[2] Rolf Adam and Leanne Bischof. ``Seeded Region Growing'',
volume 16. June 1994.
[3] John Canny. ``A Computational Approach to Edge Detection''.
November 1986.
[4] Dorin Comaniciu and Peter Meer. ``Mean Shift: A Robust Approach
Toward Feature Space Analysis'', volume 24. MAY 2002.
[5] Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest, and
Clifford Stein. ``Introduction to Algorithms''. The MIT Press, second
edition, 2001.
[6] Richard O. Duda, Peter E. Hart, and David G. Stork. ``Pattern
Classification''. Wiley-Interscience, second edition edition, 2000.
[7] Jianping Fan, Yuli Gao, and Hangzai Luo. ``Hierarchical
Classification for Automatc Image Annotation''. July 2007.
[8] Jianping Fan, Yuli Gao, Hangzai Luo, and Guangyou Xu. ``Auto-
matic Image Annotation by Using Concept-Sensitive Salient Objects
for Image Content Representation''. 2004.
[9] Pedro F. Felzenszwalb and Daniel P. Huttenlocher. ``Efficient
Graph-Based Image Segmentation'', volume 59. 2004.
[10] Rafael C. Gonzalez and Richard E. Woods. ``Digital Image
Processing''. Prentice Hall, second edition edition, 2002.
[11] Chian hung Ho. ``Modeling and Reconstruction of 3D Rat''. Chang-
Gung University, Master Thesis, July 2005.
54
[12] Ramesh Jain, Rangachar Kasturi, and Brian G. Schunck. ``Machine
Vision''. MIT Press and McGraw Inc., New York, 1995.
[13] Dzung L. Pham, Chenyang Xu, and Jerry L. Prince. ``Current
Methods in Medical Image Segmentation'', volume 2. 2000.
[14] A. Rosenfeld and A.C. Kak. ``Digital Picture Processing''.
Academic Press, second edition edition, 1982.
[15] Will Schroeder, Ken Martin, and Bill Lorensen. ``The Visualization
toolkit''. Prentice Hall, Upper Saddle Riverm New Jersey, 1998.
[16] Linda G. Shapiro and George C. Stockman. ``Computer Vision''.
Prentice-Hall, 2001.
[17] Ranjith Unnikrishnan and Caroline Pantofaru. ``Toward Objective
Evaluation of Image Segmentation Algorithms'', volume 29. JUNE
2007.
[18] Kimon P. Valavanis, Timothy Hebert, Ramesh Kolluru, and Nikos
Tsourveloudis. ``Mobile Robot Navigation in 2-D Dynamic
Environments Using an Electrostatic Potential Field'', volume 30.
MARCH 2000.
[19] Ronald E. Walole, Raymond H. Myers adn Sharon L. Myers, and
Keying Ye. ``Probability And Statistics for Engineer And Scientists''.
Prentice Hall, seventh edition edition, 2002.
[20] Shu-Yen Wan and William E. Higgins. ``Symmetric Region
Growing'', volume 12. Sep 2003.
[21] Shu-Yen Wan, Cherng-Min Ma, and Eric Nung. ``Multi-
dimensional Image Segmentation Using Seed-Invariant Region
Growing''. Taiwan, December 2001.
55
[22] Shu-Yen Wan and Eric Nung. ``Seed-inveriant Region Growing: It's
Properties and Applications to 3-D Medical CT Images''. 2001.
[23] Lei Zhang, Xun Wang, Nicholas Penwarden, and Qiang Ji. ``An
Image Segmentation Framework Based on Patch Segmentation''.
2006.