臺灣博碩士論文加值系統

近年來，主動輪廓模組影像分割技術已被廣泛地應用於臨床醫學影像處理上，主動輪廓模組自動化影像分割已被證實可有效且準確的分割高雜訊、低對比影像中之感興趣區域。臨床應用上，除分割出病變區輪廓外，如何準確地求得影像中病變區域的體積是相當重要的。然而，目前文獻中針對主動輪廓模組影像分割後之體積定量均以直接計數法加以定量，雖對部分簡單的分割區域量測之面積誤差很小，但對於分割區域有內凹或外凸之複雜形狀，其誤差則相當大。本論文之目的旨在發展一準確且有效之自動化體積定量方法，也就是邊緣連結與區域填充法，來量測主動輪廓模組分割後之病變區域，以解決直接計數法定量所遭遇的問題。本論文主要以不同類型電腦模擬影像對此自動化體積定量法進行評估其準確性，同時，探討影像矩陣對主動輪廓模組影像分割及區域定量之影響。
使用完整封閉輪廓類型之電腦模擬影像進行評估區域填充法之準確度。研究證實，區域填充法之準確度為100%，而以直接計數法量測之平均百分誤差為15.85%。
使用離散輪廓類型之電腦模擬影像進行評估邊緣連結之準確度。研究證實，邊緣連結與區域填充法之平均百分誤差為0.03%，遠小於直接計數法之平均百分誤差為15.08%。顯示邊緣連結及區域填充法執行面積定量比直接計數法更可靠且準確度更高。
使用區域類型之電腦模擬影像進行評估主動輪廓模組影像分割之準確度。以邊緣連結與區域填充法量測主動輪廓模組分割後之感興趣區域面積之平均百分誤差為7.52%，遠小於直接計數法之平均百分誤差為20.36%。顯示邊緣連結及區域填充法執行主動輪廓模組分割區域面積定量比直接計數法更可靠且準確度更高。
使用不同影像矩陣之區域類型電腦模擬影像進行評估影像矩陣對主動輪廓模組影像分割及區域定量之影響。研究證實其平均百分誤差分別為17.55%(256 × 256)，10.89%(512 × 512)及6.21%(1024 × 1024)。顯示影像矩陣越大，其平均百分誤差越小，影像分割的準確度越高。同樣地，由形態學證實影像矩陣愈大的臨床腦部病變MRI影像，其影像分割準確度愈高，體積定量的準確度也隨分割之準確度提高而升高。本論文所發展的體積定量方法，可準確定量所有經主動輪廓模組分割後之感興趣區域面積，可適用於定量不規則形狀的惡性腫瘤影像之腫瘤面積。此自動化體積定量方法將有助於提高醫師在臨床診斷上的效率、準確度、客觀性及手術前後或治療計畫的評估。

Recently, active contour model (ACM) algorithm has been widely used to perform image segmentation in many clinical imaging modalities. It has been proven to be an efficient and accurate imaging segmentation method to segment the region of interest in high noise or low contrast image. In addition to segment out the lesion contour, how to accurately quantify the volume of the lesion in an image is very important. The direct counting method has been used to quantify the volume of ACM segmented regions. Although the error of the direct counting method is quite small in some segmented regions with simple shapes, the error is quite large in most segmented region with some concave or convex shapes. In order to alleviate the problems encountered by direct counting method, a new volume quantification method were proposed.
The purpose of this study is to develop an accurate and efficient automatic volume quantification method for the ACM segmented region. The method presented herein includes edge linking and region filling processes. Various types of computer simulated images were created to evaluate the accuracy of the automatic volume quantification method. In the mean while, the relationship between image matrix size and the quantification of ACM segmented region were also investigated. These computer simulated images were categorized into three types, closed contour, discrete contour and regional types.
Images with closed contour were used to evaluate the accuracy of the region filling process. The results showed that the accuracy of region filling process was 100%, and the average percentage error of direct counting method was 15.85%. It means that region filling process is more accurate and more reliable than direct counting method in quantifying the segmented region.
Images with discrete contour were used to evaluate the accuracy of the edge linking process. The results showed that the average percentage error of edge linking and region filling process was 0.03% which is much smaller than that of direct counting method 15.08%. It means that edge linking and region filling process is more accurate and more reliable than direct counting method in quantifying the segmented region.
Images with regional type were used to evaluate the accuracy of the ACM image segmentation. The results showed that the average percentage error of edge linking and region filling process after ACM image segmentation process was 7.52% which was much smaller than that of direct counting method 20.36%. It means that edge linking and region filling process after ACM image segmentation process is more accurate and more reliable than direct counting method in quantifying the segmented region.
Regional type images with various matrices were used to evaluate the correlation between image matrix size and the quantification of ACM segmented region. The results showed that the average percentage errors of edge linking and region filling process after ACM image segmentation process were 17.55%(256 × 256), 10.89%(512 × 512) and 6.21%(1024 × 1024). It means that the larger the image matrix, the smaller the average percentage error. The larger the image matrix, the higher the accuracy of ACM image segmentation process. From the point of view of morphology, it also showed that the larger the image matrix of clinical MR images, the higher the accuracy of ACM image segmentation process and volume quantification. The proposed volume quantification method can be used in quantifying the volume of segmented region of interest after ACM image segmentation process accurately. It is capable of quantifying the volume of malignant lesion with irregular shapes. It might be useful to increase the efficiency, accuracy, externality in clinical diagnosis and the evaluations of pre-surgical, post-surgical or treatment plan.

目錄.............................................................................................................I
圖目錄......................................................................................................IV
表目錄.....................................................................................................VII
中文摘要...............................................................................................VIII
英文摘要...................................................................................................X
第一章 緒論............................................................................................1
1-1研究背景及文獻回顧...........................................................1
1-2研究動機及目的...................................................................3
1-3論文架構.........................................................................4
第二章 主動輪廓模組..........................................................................5
2-1主動輪廓模組理論背景.........................................................5
2-1.1主動輪廓模組的演進................................................5
2-1.2模組參數..................................................................6
第三章 研究方法..................................................................................9
3-1整體實驗設計...........................................................................9
3-2 程式發展.....................................................11
3-2.1.1 離散輪廓連續化的研究背景..........11
3-2.1.2 離散輪廓連續化的演算法..........................11
3-2.1.3 離散輪廓連續化的步驟.......................... 18
3-2.2.1 區域填充(Region Filling)法的理論背景..
.................................................................................20
3-2.2.2 膨脹(dilation)........................................20
3-2.2.3 區域填充演算法..........................................21
3-2.2.4 區域填充的步驟...................................... 21
3-3影像取得.............................................................................25
3-3.1 電腦模擬影像取得.................................................25
3-3.1.1輪廓類型之電腦模擬影像.........................26
3-3.1.2區域類型之電腦模擬影像.........................28
3-3.2 MRI影像取得...............................................30
3-4 影像矩陣對主動輪廓模組進行影像分割之準確度評估...............................................30
第四章 結果..........................................................................................33
4-1電腦模擬影像實驗結果.....................................................33
4-1.1 輪廓類型之電腦模擬影像的實驗結果..........33
4-1.1.1完整封閉輪廓之電腦模擬影像的實驗結果.............................34
4-1.1.2離散輪廓之電腦模擬影像的實驗結果..............................................................38
4-1.2 區域類型之模擬影像的實驗結果..........................42
4-2影像矩陣對主動輪廓模組進行影像分割之準確度評估結果.............................................49
4-2.1區域類型電腦模擬影像...........................49
4-2.2臨床腦部病變MRI影像實驗結果.........................57
第五章 討論與結論..............................................................................60
參考文獻..................................................................................................64

圖目錄
圖3-1 實驗流程圖。................................................................................10
圖3-2 第一類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 時，其中心點位置不進行補點的所有情形。.......12
圖3-3 第二類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 時，其中心點位置不進行補點的所有情形。.........13
圖3-4 第三類類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 且兩邊界點之中心點與矩陣中心點的歐基里德距離為 時在其中心點位置進行補點的所有情形。...........13
圖3-5 第四類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 且兩邊界點之中心點與矩陣中心點的歐基里德距離為 時在其中心點位置進行補點的所有情形。................14
圖3-6 第五類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 且兩邊界點之中心點與矩陣中心點的歐基里德距離為0時在其中心點位置進行補點的所有情形。.....................14
圖3-7 第六類型，3×3矩陣內存在兩個邊界點且兩邊界點間之歐基里德距離為 且兩邊界點之中心點與矩陣中心點的歐基里德距離為1時在其中心點位置不進行補點的所有情形。...................15
圖3-8 第一類型，3×3矩陣內存在三個邊界點且三邊界點間之距離分別為1、2、 時，在其中心點位置進行補點的所有情形。.......16
圖3-9 第二類型，3×3矩陣內存在三個邊界點且三邊界點間之距離分別為1、 、 時，在其中心點位置進行補點的所有情形。...........................17
圖3-10 3×3矩陣內存在三個邊界點且在其中心位置填補斷點之類別。....................................................18
圖3-11 離散輪廓連續化之流程圖。.............................................19
圖3-12 種子在封閉輪廓內之位置。.................................................22
圖3-13 結構元素。.....................................................22
圖3-14區域填充的過程。.........................................................................23
圖3-15 區域填充之流程圖。....................................................................24
圖3-16 基本類型之電腦模擬影像。........................................................26
圖3-17複雜類型之電腦模擬影像。..............................................26
圖3-18 完整封閉輪廓之電腦模擬影像實驗流程圖。.....................................27
圖3-19 離散輪廓電腦模擬影像之實驗流程圖。....................................28
圖3-20區域類型-基本類型之電腦模擬影像。........................................29
圖3-21區域類型-複雜類型之電腦模擬影像。........................................29
圖3-22 腦部軸狀切面T1加權影像。......................................................30
圖3-23其他類型之模擬影像。......................................................31
圖4-1完整封閉輪廓之電腦模擬影像之計數圖示。...............................35
圖4-2離散輪廓之電腦模擬影像之計數圖示。.......................................39
圖4-3 區域類型之電腦模擬影像分割及面積定量計數圖示。..............44
圖4-4區域類型之電腦模擬影像分割及面積定量計數圖示。...............46
圖4-5區域類型電腦模擬影像之主動輪廓模組分割及面積定量影像。..................................................51
圖4-6區域類型電腦模擬影像之主動輪廓模組分割及面積定量影像。..............................................................................................52
圖4-7區域類型電腦模擬影像之主動輪廓模組分割及面積定量影像。..............................................................................................53
圖4-8區域類型電腦模擬影像之主動輪廓模組分割及面積定量影像。..............................................................................................54
圖4-9區域類型電腦模擬影像之主動輪廓模組分割及面積定量影像。..............................................................................................55
圖4-10為三種不同矩陣電腦模擬影像平均百分誤差之關係圖。.....57
圖4-11臨床腦部病變MRI影像實驗。.................................................58

表目錄
表4-1完整封閉輪廓之電腦模擬影像實驗。...........................................37
表4-2離散輪廓之電腦模擬影像實驗。….............................................41
表4-3離散輪廓之電腦模擬影像獨立樣本t檢定之統計。.....................42
表4-4 區域類型之電腦模擬影像經主動輪廓模組分割之定量。........48
表4-5 區域類型之電腦模擬影像獨立樣本t檢定之統計。...................48
表4-6 主動輪廓模組對不同影像矩陣實驗。...................................56
表5-1不同類型電腦模擬影像之平均百分誤差。...................................61
表5-2不同影像矩陣分割影像之平均百分誤差。.................................. 62

【1】Haralick R and Shapiro L, “Image segmentation techniques”, Cornput. Vision, Graphics, Image Proc 1985; 29c 100-132.
【2】Saha P K, Udupa J K, “Optimum image thresholding via class uncertainty and region homogeneity”, Pattern Analysis and Machine Intelligence, IEEE Transactions on 2001; 23: 689- 706.
【3】Julio C G, Catherine K, Sharmila M, and Nola H, “Application of a fuzzy inference system to the quantification of 3D magnetic resonance imaging of breast tissue”, Medical Imaging, 2004: Image Processing 2004; 5370 :1539-1547.
【4】Cvancarova M, Albregtsen F, Brabrand K, Samset E, “Segmentation of ultrasound images of liver tumors applying snake algorithms and GVF”, International Congress Series 2005; 1281: 218-223.
【5】陳淑嬌,梯度向量流主動輪廓模組於抹片細胞邊界偵測之應用, 國立中興大學電機工程研究所碩士論文﹐台灣﹐2003.
【6】Davatzikos C A and Prince J L, “An active contour model for mapping the cortex” , IEEE Trans Med. Image 1995; 14: 65-80.
【7】Xu C, Pham D L, Rettmann M E, Yu D N, and Prince J L, “ Reconstruction of human cerebral cortex from magnetic resonance images”, IEEE Trans Med. Image 1999; 18: 467-480.
【8】Kovalski G, Beyar R, Shofti R, and Azhari H, “Three-dimensional automatic quantitative analysis of intravascular ultrasound images” , Ultrasound Med. Biol. 2000; 26: 527-537.
【9】Xu C and Prince J L, “Snakes, shapes, and gradient vector flow”, IEEE Trans Image Proc. 1998; 7: 359-369.
【10】Monga O, Derichc R, Malandain G, arid Coccpcrcz J P, “Recursive filtering and edge tracking: Two primary tools or 3D edge detection”, huge Vis. Cornput. 1991; 9 :203-214.
【11】Jing F, Li M, Zhang L, Zhang H J, Zhang B, “Learning in Region- Based Lmagc: Retrieval”, Proc. International Confercnce on Image and Video Retricval 2003.
【12】Chuang C H, Lie W N, “Region growing based on extended gradient vector flow field model for multiple objects segmentation”, Image Processing, Proceedings. International Conference 2001; 3: 74-77.
【13】Alakuijala, J, Laitinen, J, Sallinen, S, Helminen, H, “New efficient image segmentation algorithm: competitive region growing of initial regions”, Engineering in Medicine and Biology Society, 1995. IEEE 17th Annual Conference 1: 409-410.
【14】Marte O, Marais P “Model-based Segmentation of CT Images”, Proceedings of SAICSIT 2001; 96 - 101.
【15】Lin P, Zheng C X, Yang Y, “Model-based medical image segmentation: a level set approach”, Intelligent Control and Automation, 2004. WCICA 2004. Fifth World Congress, 2004; 6: 5541-5544.
【16】Li X, Hamarneh G, “Modeling Prior Shape and Appearance Knowledge in Watershed Segmentation”, Proceedings of the Second Canadian Conference on Computer and Robot Vision (CRV’05).
【17】Metaxas D, Ting C, “A hybrid 3D segmentation framework”, Biomedical Imaging: Macro to Nano, 2004. IEEE International ymposium ; 2004; 1: 13-16.
【18】Li H, Yezzi A, “A Hybrid Medical Image Segmentation Approach Based on Dual-Front Evolution Model”, Image Processing, 2005. ICIP 2005. IEEE International Conference on 2: 810-813.
【19】Gauch J M, Pien H, Shah J, “Hybrid deformable models for three-dimensional biomedical image segmentation”, Nuclear Science Symposium and Medical Imaging Conference., 1994 IEEE Conference Record 4: 1935-1939.
【20】Kass M, Witkin A, and Terzopoulos D, “Snakes: Active Contour models” , Int. J. Comput. Vision 1988;1: 321-331.
【21】Cohen L D, “On active contour models and balloons,” CVGIP: Imag Understanding, 1991; 53: 211-218.
【22】Czerwinski R N, Jones D L, O'Brien W D, “Line and Boundary Detection in Speckle Images”, Jr.; Image Processing, IEEE Trans.1998; 7: 1700-1714.
【23】Czerwinski R N , Jones D L, O'Brien W D, “Detection of lines and boundaries in speckle images-application to medical ultrasound”, Jr.;Medical Imaging, IEEE Trans. 1999; 18: 126-136.
【24】Chang R F, Wu W J, Moon W K, Chen W M, Lee W, Chen D R, “Segmentation of breast tumor in three-dimensional ultrasound images using three- dimensional discrete active contour model”, Ultrasound in Medicine and Biology 2003; 29: 1571-1581.
【25】Gonzalez R F and Wintz P, “Digital Image Processing, 3rd Ed.” Addison-Wesley 1992.