臺灣博碩士論文加值系統

在這篇論文中，我們提出一個更有效率且更準確的方法來切割肺部的斷層掃描影像和偵測肺部腫瘤。首先，我們利用thresholding、morphology closing和labeling等影像處理的技術來切割肺葉影像並從中擷取可疑的腫瘤區域。然後，我們從可疑的腫瘤區域中擷取三個特徵值，分別是面積、圓滑度和平均亮度。最後利用這三個特徵值和我們架構的Fuzzy Inference Rules和Neural Fuzzy系統來辨別腫瘤。
我們測試的樣本有29個臨床上的病例，總共有583張512x512x8位元的影像。Fuzzy Inference Rules和Neural Fuzzy系統整體的正確偵測率皆可達89.3%，而偽陽率平均每張影像分別為0.3及0.21，由這些數據可知本系統增進了偵測率並降低了偽陽率，而在醫療臨床的實行上也有相當的可行性。

In this thesis, we present new integrated methods to segment the lung CT images and to detect the nodules more efficiently and correctly. We employed series of image processing techniques including thersholding, morphology closing, and labeling to segment the lung area and obtain the region of interests (ROIs). We extract three main features, circularity, size of area, and mean brightness from ROIs and identify the nodules with pure fuzzy inference rules and neural fuzzy model. Twenty-nine clinical cases involving 583 images (512X512X8 bits) were tested in our study. The detection rate of the proposed methods are 89.3%, and the false positives are approximately 0.3 and 0.21 per image respectively. This result demonstrates that our method improves the detection rate and reduces false positive compaired to other approaches. The studies have shown high potential implementation of this system in clinical practice.

List of Figures iii
List of Tables v
Abstract(Chinese) vi
Abstract(English) vi
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . 1
1.2 Methodology .. . . . . . . . . . . . . . . . . . . 2
1.3 Guide of the Thesis . . . . . .. . . . . . . . . . . 3
2 Literature Survey. . . . . . . . . . . . . . . . . . . 4
2.1 Image Segmentation . . . . . . . . . . . . . . . 4
2.2 Lung Nodules Detection Techniques . .. . . . . . . . 7
3 Lung Nodule Segmentation and Feature Extraction 16
3.1 System Architecture . . . . . . . . . . . . . . 16
3.2 Segmentation of Lung Field . . . . . . . . . . . 18
3.3 Region of Interest (ROI) Selection . . . . . . . 21
3.4 Feature Extraction . . . . . . . . . . . . . . . 21
4 Neural Fuzzy Identification for True Nodule 25
4.1 Fuzzy Inference System . . . . . . . . . . . . . 27
4.2 Neural FuzzyModel . . . . . . . . . . . . . . . . 30
5 Experiment Results 45
5.1 System Performance and Comparison . . . . . . . . 45
5.2 Evaluation of False Positive . . . . . . . . . . 50
5.3 Evaluation of False Negative . . . . . . . . . . 50
6 Conclusions & Future Work 54
6.1 Conclusions . . . . . . . . . . . . . . . . . . . 54
6.2 FutureWork . . . . . . . . . . . . . . . . . . . 55

[1] S. Armato, M. Giger, and H. MacMahon. Automated lung segmentation in digitized posteroanterior chest radiographs. Academic Radiology, 4:245—255, 1998.
[2] M. S. Brown, M. F. McNitt-Gray, N. J. Mankovich, J. G. Goldin, J. Hiller, L. S. Wilson, and D. R. Aberle. Method for segmenting chest ct image data using an anatomical model: Preliminary results. IEEE Transactions on Medical Imaging, 16:828—839, December 1997.
[3] F. Carrascal, J. Carreira, M. Souto, P. Tahoces, L. Gomez, and J. Vidal. Automatic calculation of total lung capacity from automatically traced lung boundaries in posteroanterior and lateral digital chest radiographs. Medical Physics, 25:1118—1131, 1998.
[4] J. Duryea and J. Boone. A fully automatic algorithm for the segmentation of lung elds in digital chest radiographic images. Medical Physics, 22:183—191, 1995.
[5] Jun-Feng Guo, Yuan-Long Cai, and Yu-Ping Wang. Morphology-based interpolation for 3d medical image reconstruction. Computerized Medical Imaging and Graphics, 19:267—279, 1995.
[6] C. I. Henschke, D. I. McCauley, D. F. Yanlkelevitz, D. P. Naidich, G. McGuinness, O. S. Miettinen, D. M. Libby, M. W. Pasmantier, J. Koizumi, N. K. Altorki, and J.P. Smith. Early lung cancer action project : overall design and ndings from baseline screening. The LANCET, 354:99—105, July 1999.
[7] Shiying Hu and Eric A. Homan. Automic lung segmentation for accurate quantitation of volumetric x-ray ct images. IEEE Transactions on Medical Imaging, 20:490—498, June 2001.
[8] K. Kanazawa, Y. Kawata, and N. Niki. Computer-aided diagnosis for pulmonary nodules based on helical ct images. Computerized Medical Image and Graphics, pages 157—167, 1998.
[9] Y. Kato, K. Yamada, F. Oshita, I. Nomura, K. Noda, T. Yamagata, and M. Tajiri. Helical thin-section ct high-resolution image analysis of resected peripheral adenocarcinomas of the lung less than 1 cm in diameter. Lung Cancer, 18:214, August 1997.
[10] Y. Kawata, N. Niki, H. Ohmatsu, R. Kakinuma, K. Eguchi, and
N. Moriyama. Quantitative surface characterization of pulmonary nodules based on thin-section ct images. IEEE Trans. Nuclear Science, 45:2132—2138, 1998.
[11] M. Kubo, N. Niki, S. Nakagawa, K. Eguchi, and M. Kaneko. Extraction algorithm of pulmonary ssures from thin-section ct images based on linear feature detector method. IEEE Transactions on Nuclear science, 46:2128—2133, December 1999.
[12] Yongbum Lee and Takeshi Hara. Automated detection of pulmonary nodules in helical ct images based on an improved template-matching technique. IEEE Transactions on Medical Image, pages 557—563, July 2001.
[13] Chin-Teng Lin and C.S. George Lee. Neural Fuzzy Systems. Prentice Hall P T R, New York, 1996.
[14] Jyh-Shyan Lin and Panos A. Ligomenides. A hybrid neural digital computer-aided diagnosis system for lung nodule detection on digitized chest radiographs. Seventh Annual IEEE Symposium on Computer-Based Medical System, pages 207—212, 1994.
[15] T. Okumura, T. Miwa, J. Kato, S. Yamamoto, M. Matsumoto,
Y. Tateno, T. Iinuma, and T. Matsumoto. Variable n-quoit lter applied for automatic detection of lung cancer by x-ray ct. in Proc. CAR’98, pages 242—247, 1998.
[16] M.G. Penedo, M.J. Carreira, A.Mosquera, and D. Cabello. Computeraided diagnosis: A neural-network-based approach to lung nodule detection. IEEE Transactions on Medical Imaging, 17:999, December 1998.
[17] T. Tozaki, Y. Kawata, N. Niki, H. Ohmatsu, and R. Kakinuma. Pulmonary organs analysis for dierential diagnosis based on thoracic thinsection ct images. International conference on image processing, pages 1332—1336, 1998.
[18] Bram van Ginneken, Bart M. ter Haar Romeny, and Max A. Viergever. Computer-aided diagnosis in chest radiography: A survey. IEEE Transactions on Medical Image, 20:1228—1241, December 2001.
[19] X.W. Xu, S. Katsuragawa, K. Ashizawa, H. MacMahon, and K.Doi. Analysis of image features of histograms of edge gradient for false positive reduction in lung nodule detection in chest radiographs. Proc SPIE, page 3338, 1998.