臺灣博碩士論文加值系統

在臨床診斷上，經由造影技術所獲得的醫學影像資料，均以沖印成膠片的方式閱覽及儲存。但是，往往會因為環境溫度的改變，乃至於擺放方式的不適當，使得膠片變形或起化學變化，進而造成膠片上的影像扭曲或失真。基於保存資料的完整性及加速電子化作業，透過掃描器的使用，將膠片影像自動轉換成數位化資料檔案並加以儲存有其實務性及必要性。
本研究主要目的，是在發展一套快速且自動的方法，能將一張臨床上含有多個影像區塊的膠片(film)格式，結合現今的電腦科學、光學設備儀器及影像處理技術，讓底片上的影像區塊自動轉換成多個獨立的數位化影像檔，這將極有利於後續各項電腦化作業及數據分析處理。
本研究使用projection profile的方法，對數位化影像做水平（及垂直）投影分佈，並進行區塊影像分割。但經由掃描器數位化操作後的影像，極容易成為歪斜不正的區塊影像，而無法產生明顯的峰谷值 (valleys)，導致分割失敗，故本研究計劃以逐角偵測及哈福轉換直線偵測為前置步驟，自動偵測出區塊影像與水平軸之夾角，獲得全幅影像之偏斜程度，進行影像校正。
以本文所提之方法，利用投影量逐角偵測法，以0.01度為單位來校正影像，經過實驗模擬及醫學影像實做結果，校正後能更精確的切割區塊影像。

Abstract:
The medical image data captured from image manufacture technique , is always printed as filmy image for reading and storage on clinical diagnosis. However, the environment temperature change and bad storeroom cause chemical change, though films are easier to store. Through the scanner, we turn film image into digitalized data storage, for data completion and data digitalization.
This paper presents a speedy way to combine a personal computer with optical peripheral equipment and image processing. It can automatically transfer a clinical film that with many block images into many independent digital image. It is very useful for data computerization and analysis.
We use projection profile to segment the block image in the vertical and horizontal projection of digitalized image data. It is easy to create skew block image and fails to segment image, through the operation of scanner digitalization. We take Hough Transform’s line detect as the first process, for auto detection of the angle between vertical and horizontal and get the degree of the scan digitalized image in this research.
In this paper, detect utilize projection profile is used to angle by angle, calibrate image in 0.01 degree. It can precisely segment block image as the result of simulation and medical image auto-conversion.

目錄
中文摘要……………………………………………………………Ⅰ
英文摘要……………………………………………………………Ⅲ
誌謝…………………………………………………………………Ⅳ
目錄…………………………………………………………………Ⅴ
圖表目錄……………………………………………………………Ⅵ
一、緒論
1.1 前言………………………………………………………1-1
1.2 相關研究…………………………………………………1-2
1.3 論文架構…………………………………………………1-3
二、膠片影像資料數位化
2.1 醫學膠片影像……………………………………………2-1
2.2 醫學膠片影像數位化……………………………………2-4
三、數位影像校正及區塊影像切割方法
3.1 去雜訊……………………………………………………3-2
3.2 運用邊緣線尋找傾斜角度………………………………3-4
3.2.1 影像膨脹……………………………………………………3-5
3.2.2 膨脹影像邊緣線偵測………………………………3-6
3.2.3 哈福轉換(Hough Transform)………………………3-9
3.3 逐角偵測……………………………………………………3-10
3.4 影像校正……………………………………………………3-13
3.5 影像切割…………………………………………………3-15
四、實驗結果與分析
4.1 模擬驗證與分析…………………………………………4-1
4.2 實際醫學影像切割………………………………………4-5
4.2.1 電腦斷層(CT)膠片影像……………………………4-5
4.2.2 核磁共振(MRI)膠片影像…………………………4-7
五、結論
六、參考文獻

參考文獻[1] Chiu L. Yu , Yuan Y. Tang and Ching Y. Suen , “Document Skew Detection Base on the Fractal and Least Squares Method, ”in IEEE , 1995, pp1149-1152.[2] Yasuto ISHITINI , ”Document Skew Detection Based on Local Region Complexity, ”in IEEE , 1993, pp49-52.[3] Reza Safabakhah and Shahram Khadivi , “Document Skew Dection Using Minimum -Area Bounding Rectangle, ”in IEEE , 2000, pp253-258.[4] S C.H,J L.F , and D P.D , “A Document Skew Detection Method Using Run-Length Encoding and Hough Transform, ”in IEEE , pp464-468 , 1990.[5] Younki Min , S-B C , Yillbyung Lee , “A Data Reduction Method for Efficient Document Skew Estimation Based on Hough Transformation” , IEEE , 1996 pp732-736 ,.[6] Huei-Fen Jiang , Chin-Chuan Han , and Kuo-Chin Fan , “A Fast Approach to Detect and Correct Skew Documents, ”in Proceedings of ICPR ‘96 , 1996 pp230-233 ,.[7] C. Sun , D. Si , “Skew and Slant Correction for Document Images Using Generient Direction, ”in IEEE , 1997 ,pp142-146.[8] E.K, N.F, G.K , “New Algorithms For Skewing Correction And Slant Removal on Word-Level, ”in IEEE , 1999,pp1159-1162.[9] Kazuhide Sugawara , “Weighted Hough Transform on a Gridded Image Plane, ”in IEEE , 1997,pp701-704.[10] Mahbubul Islam Chowdhury , John A Robinson , “Improving image segmentation using edge information, ”in IEEE , 2000 , pp312-316.[11] Wendy J.H and Matrin J.R , “Modeling Speech Variability With Segmental HMMS , “in IEEE , 1996 , pp447-450.[12] Anniwear YMIN , Yoshinao AOKI, “On the Segmentation of Multi-Font Printed Uygur Scripts, “in IEEE , Proceedings of ICRP’96 , pp215-219.[13] Sung-Il Chien and Yung-Mok Bael, “Hierarchical Block Matching Method for Fast Rotation of Binary Images, “in IEEE Transactions on Image Processing, Vol. 10,No.3,March 2001, pp483-489.[14] Kaushik Mahata and A.G. Ramakrishnan, Serior Member IEEE, “A Novel Scheme For Image Rotation Document Processing, “in IEEE , 2000 , pp594-596.[15] Dimitrios Drivas and Adnam Amin , “Page Segmentation and Classification Utilising , ”in IEEE , 1995 , pp610-614.[16] A Antronacopoulos and R T Ritchings, “Segmentation And Classification of document images, “in 1995 The Institution of Electrical Engineers. Printed and published by the IEE, Savoy Place, London Wc2R OBL, UK, pp1-16.[17] C. M. Leung and W.-S. Lu, “A Two-Step Regularization Approach for Image Noise Removal,” IEEE, 1995.[18] Artur Stankiewicz*, Idzi Merta, Leszek R. Jaroszewicz, “The Edge Regularization of Noise Image Method, ” Fourth IEEE International Workshop on Cellular Networks and their Application, Seville, Spain, June 24-26,1996.[19] C. J. Dykstra, R. Harrop, “Automatic Estimation and Removal of Emission Image Noise,” IEEE 1998.[20] Gonzalez, R.C and Woods , R.E., 1992. Digital Image Processing Addison-Wesley Publishing Company.[21] 連國珍，”數位影像處理”，儒林圖書公司。[22] Milan Sonka, Vacklaw Hlavac, Roger Boyle, “Image Processing, Analysis, and Machine Vision,” Second Edition, 1998.[23] EARL Gose, Richard Johnsonbaugh, Steve Jost, “Pattern Recognition and Image analysis,” 1996.[24] Robert J. Schalkoff, Clemson University, “Pattern Recognition : Statistical, Structural and Neural Approaches,” 1992.[25] Kuo-Lung Ku, “Algorithm and Architectures for the Hough Transformation, 1987.