臺灣博碩士論文加值系統

隨著電腦斷層攝影技術不斷地改良，普遍地用來診斷腎臟病變的腹腔電腦斷層影像張數也隨之大量地增加，造成閱片者重大的負擔。本論文提出一種以紋理學習為基礎之三維腎實質分割系統應用於腹腔電腦斷層影像。此系統結合紋理學習以及區域內的均勻性為基礎的兩種方法，設計用於腎實質的自動描繪。本論文所使用的紋理學習法，是採用灰階共生矩陣之特徵與類神經網路去辨識電腦斷層影像中的各像素是否為腎實質。本論文所使用的區域內的均勻性方法是結合分割單張電腦斷層影像內腎實質的二維區域成長法以及會傳遞分割結果到鄰近電腦斷層影像的三維區域成長法。此區域成長的成長標準是由類神經網路去評估區域的均勻性決定是否應該要成長。本系統以測試十組的腹腔斷層掃描影像來評估其效能，比較手動分割的結果與自動分割的結果之間的DICE相似係數。經過十組腹腔斷層掃描影像測試後，本系統達到平均0.87之DICE相似係數，此結果顯示了兩種分割結果是非常相近地。總結而言，我們的系統能應用於描繪腎實質之應用程式或是做為電腦輔助之腎臟病變診斷系統的前處理器。

Abdominal CT images are commonly used for the diagnosis of kidney diseases. With the advances of CT technology, processing of CT images has become a challenging task mainly because of the large number of CT images being studied. This paper presents a texture-learning based system for the three-dimensional (3D) segmentation of renal parenchyma in abdominal CT images. The system is designed to automatically delineate renal parenchyma and is based on the texture-learning and the region-homogeneity-based approaches. The first approach is achieved with the texture analysis using the gray-level co-occurrence matrix (GLCM) features and an artificial neural network (ANN) to determine if a pixel in the CT image is likely to fall within the renal parenchyma. The second approach incorporates a two-dimensional (2D) region growing to segment renal parenchyma in single CT image slice and a 3D region growing to propagate the segmentation results to neighboring CT image slices. The criterion for the region growing is a test of region-homogeneity which is defined by examining the ANN outputs. In system evaluation, 10 abdominal CT image sets were used. Automatic segmentation results were compared with manually segmentation results using the Dice similarity coefficient. Among the 10 CT image sets, our
system has achieved an average Dice similarity coefficient of 0.87 that clearly
shows a high correlation between the two segmentation results. Ultimately, our system could be incorporated in applications for the delineation of renal parenchyma or as a preprocessing in a CAD system of kidney diseases.

Contents
Abstract in Chinese ............................................I
Abstract ......................................................II
Acknowledgement................................................IV
Contents .......................................................V
List of Figures ..............................................VII
List of Tables ..............................................VIII
1. INTRODUCTION ...........................................1
2. PREPROCESSING ..........................................5
2.1 Image Conversion .......................................5
2.2 Spine Localization .....................................7
2.3 Finding a Seed for Each Kidney .........................8
3. ARTIFICIAL NEURAL NETWORK TRAINING ....................10
3.1 Texture Feature Extraction ............................10
3.2 ANN Training ..........................................12
4. SEGMENTATION OF RENAL PARENCHYMA ......................14
5. RESULTS ...............................................17
5.1 CT Image Database .....................................17
5.2 ANN Training ..........................................17
5.3 Evaluation of Renal Parenchyma Segmentation ...........18
6. CONCLUSION ............................................23
References ....................................................25
List of Figures
Fig.1. A simplified flow chart of our texture-learning based system for 3D segmentation of renal parenchyma in abdominal CT images.........................4
Fig.2. An example of the spine and kidney localization where (a) is the resulting original CT image; (b) is the image after thresholding; (c) is the resulting image after morphological closing; and (d) is the identified spine and initial contour of renal parenchyma.......................................9
Fig.3. (a) An abdominal CT image slice for the ANN training; (b) The mask image for (a) where the red region are the renal parenchyma and the green regions are the non-renal parenchyma..............................................................................................................................14
Fig.4. Flow chart of 2D region growing in our system......................................................................................................15
Fig.5 Flow chart of 3D region growing in our system.......................................................................................................16
Fig.6. Accuracy (%) for various window size s = 3 to 15 using the renal parenchyma and non-renal parenchyma given in Fig.3................................18
Fig.7. An example of the 3D segmentation of renal parenchyma in abdominal CT images.......................................................................19
List of Tables
Table.1. Dice similarity coefficients between the manual and automatic segmentations used for system evaluation using the 10 CT image sets. Within each CT image set, the mean and the standard deviation of all the Dice similarity coefficients of all single CT image slices are given............................21
Table.2. Processing time required in the 3D segmentation of renal parenchyma in the 10 CT image sets (in seconds).........................................23

REFERENCES
[1]National Cancer Institute, U.S.A., http://seer.cancer.gov, July, 2008.
[2]Chen, E. L., Chung, P. C., Chen, C. L., Tsai, H. M. and Chang, C. I., "An Automatic Diagnostic System for CT Liver Image Classification," IEEE Trans. on Biomedical engineering, vol 45, no. 6, 783-794 (1998).
[3]Kim, H., Honda S. I., Ishikawa, Okuda, S. K., Yamamoto, A., Katsuragawa, S. and Viergever, M. A., "Automatic extraction of liver regions and classification of liver cirrhosis from the abdominal CT images," Industrial Electronics Society (IECON), 30th Annual Conference of IEEE, vol 2, 1086-1089 (2004).
[4]Lu, R., Marziliano, P. and Thng, C. H., "Liver Tumor Volume Estimation by Semi-Automatic Segmentation Method, " Engineering in Medicine and Biology Society (IEEE-EMBS), 27th Annual International Conference of the, 3296-3299 (2005).
[5]Gletsos, M., et al., "A computer-aided diagnostic system to characterize CT focal liver lesions: design and optimization of a neural network classifier, " IEEE Trans. Inf. Tech. Bio., vol. 7, no. 3, 153-162 (2003).
[6]Tsagaan, B., Shimizu, A., Kobatake, H., Miyakawa, K. and Hanzawa, Y., "Segmentation of Kidney by Using a Deformable Model," International Conference on Image Processing (ICIP), vol 3, 1059-1062 (2001).
[7]Touhami, W., Boukerroui D. and Cocquerez, J., "A Statistical Approach for Automatic Kidneys Detection," IEEE Intertional Conference on Image Processing (ICIP), vol 3, III -740-3 (2005) .
[8]Zhang, J., Zhu H. and Qian, X., "Genetic Algorithm for Edge Extraction of Glomerulus Area," Proceedings of International Conference on Information Acquisition, 335-338 (2004).
[9]Lin, D., Lei, C. and Hung, S., "Computer-Aided Kidney Segmentation on Abdominal CT Images," IEEE Trans. on Information Technology in Biomedicine, vol 10, issue 1, 59-65 (2006).
[10]Kim, S. H., Yoo, S. W., Kim, S. J., Kim, J. C. and Park, J. W., "Segmentation of Kidney without using Contrast Medium on Abdominal CT Image," Proceedings of 5th International Conference on Signal Processing (WCCC-ICSP), vol 2, 1147-1152 (2000).
[11]Kubota Y. , Mitsukura Y. , Fukumi M. , Akamatsu N. and Yasutomo M., "Automatic extraction of a kidney region by using the Q-learning," Proceedings of International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS), 536 - 540 (2004).
[12]Petrou, N. and Sevilla, P. G., [Image Processing Dealing with Texture], John Wiley & Sons Ltd, England, 275-294 (2006).
[13]Gonzalez, R. C., Woods,. R E., [Digital Image Processing], Prentice Hall (2002).
[14]Baraldi, A. and Parmiggiani, F., "An investigation of the Textural Characteristics Associated with Gray Level Co-occurrence Matrix Statistical Parameters," IEEE Trans. on Geoscience and Remote Sensing, vol. 33, no. 2, 293-304 (1995).
[15]Haralick, R. M., Shanmugam, K., and Dinstein, I., "Textural features for image classification," IEEE Trans. on Systems, Man, and Cybernetics, vo l. SMC-3, no. 6, 610-621 (1973).
[16]Haralick, R. M., and Shapiro, L. G., [Computer and Robot Vision Volume I], Addison-Wesley Publishing Company Inc., U.S.A. (1992).