跳到主要內容

臺灣博碩士論文加值系統

(18.97.14.81) 您好!臺灣時間:2025/02/13 08:43
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:沈予涵
研究生(外文):Yu-Hang Shen
論文名稱:基於卷積神經網絡之乳房彈性超音波影像電腦輔助診斷
論文名稱(外文):Computer-Aided Tumor Diagnosis for Breast Elastography Based on Convolutional Neural Network
指導教授:張瑞峰張瑞峰引用關係
指導教授(外文):Ruey-Feng Chang
口試日期:2017-07-19
學位類別:碩士
校院名稱:國立臺灣大學
系所名稱:生醫電子與資訊學研究所
學門:工程學門
學類:生醫工程學類
論文種類:學術論文
論文出版年:2017
畢業學年度:105
語文別:英文
論文頁數:37
中文關鍵詞:乳癌乳房彈性超音波電腦輔助診斷卷積神經網絡
外文關鍵詞:Breast CancerElastographyComputer-Aided Diagnosis (CADx)Convolutional Neural Network (CNN)
相關次數:
  • 被引用被引用:0
  • 點閱點閱:375
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
近年來,乳癌一直是女性癌症的主要死因之一。早期發現與早期治療是提高乳癌存活率最重要的手段之一。許多研究顯示,乳房彈性超音波(Breast Elastography)的乳癌檢出率,較傳統 B-mode 超音波檢查的診斷檢出率要高出許多。傳統電腦輔助診斷的方法,通常需要先做腫瘤影像切割,並且使用人工定義的特徵做腫瘤分類。近年來,隨著電腦硬體技術的革新,卷積神經網絡(Convolutional Neural Network, CNN)正積極的被應用在各個領域中,其中最重要的應用即為影像辨識。使用卷積神經網絡做醫學影像辨識的研究也陸續被提出,卷積神經網絡與傳統方法相比,優勢在於不須先做影像切割,及特徵是經由神經網絡自動學習擷取,可以更廣泛的應用在各種影像上,並且有效避免傳統方法對資料過擬合(Data Overfitting)的問題。本研究主要的目的是利用卷積神經網絡進行乳房彈性超音波影像的電腦輔助診斷。首先,在 B-mode 影像上選定腫瘤區域,並在彈性影像上擷取與之對應的腫瘤區域,將這兩個腫瘤區域輸入由本研究提出的卷積神經網絡架構中進行特徵學習和擷取,最後以學習到的特徵診斷腫瘤的良惡性。本研究的實驗以151個經過病理驗證的病例進行測試,包含89個良性以及62個惡性的病例。經由實驗結果,當結合B-mode與彈性影像訓練卷積神經網絡時,診斷的準確率為85.43% (129/151),靈敏性80.65% (50/62),特異性 88.76% (79/89),以及ROC曲線面積0.9202。因此可知,乳房彈性超音波結合卷積神經網絡可以有效的診斷腫瘤良惡性。
Breast cancer has become one of the most frequently diagnosed cancers and one of the major causes of cancer death in women. Early detection and treatment are the most effective way to reduce the mortality rate. Recently, many studies have shown that compared with using only conventional B-mode ultrasound, the examination with elastography can improve the diagnostic performance. The conventional computer-aided diagnosis (CADx) methods usually require tumor segmentation, and use hand-crafted features to classify the tumors. For the past few years, due to the innovation of graphics processing unit (GPU) technique, the convolutional neural network (CNN) continues to thrive. Nowadays, CNN has been applied to various fields, especially in computer vision, and has rapidly become a methodology of choice for analyzing medical images. Compared to conventional CADx methods, CNN does not need to segment the tumor manually and can learn the features automatically. CNN can also apply to various image types, and can prevent the overfitting problem in conventional methods. In this study, images of 151 biopsy-proved sonoelastography cases composed of 89 benign and 62 malignant cases are used to evaluate the diagnosis performance. According to the experimental results, the performance of utilizing both B-mode images and elastography images to train the proposed architecture achieved an accuracy of 85.43% (129/151), a sensitivity of 80.65% (50/62), a specificity of 88.76% (79/89), and an AUC value of 0.9202. This thus verifies the feasibility of utilizing breast elastography together with CNN for effective classification of breast tumors.
口試委員會審定書 .......................................... i
致謝 ...................................................... ii
摘要 ...................................................... iii
Abstract .................................................. v
Table of Contents ......................................... vii
List of Figures ........................................... viii
List of Tables ............................................ ix
Chapter 1 Introduction .................................... 1
Chapter 2 Materials ....................................... 3
2.1 Data Acquisition .................................. 3
2.1 Patients and Lesion Characters .................... 4
Chapter 3 Methods ......................................... 5
3.1 Convolutional Neural Network Architecture ......... 6
3.1.1 Convolutional Layers ........................ 7
3.1.2 Nonlinearity Layers ......................... 8
3.1.3 Max pooling Layers .......................... 9
3.1.4 Fully Connected Layers ...................... 9
3.2 Experimental Setup ................................ 10
3.3 Data Preparation .................................. 11
3.3.1 Data Preprocessing .......................... 11
3.3.1.1 B-mode images ......................... 11
3.3.1.2 Elastography .......................... 12
3.3.2 Data Augmentation ........................... 14
3.4 Testing............................................ 15
3.5 Experiment Environment ............................ 16
Chapter 4 Results and Discussions ......................... 17
4.1 Receiver Operating Characteristic (ROC) Curve ..... 17
4.2 Classification Performance ........................ 20
4.3 Statistical analysis .............................. 22
4.4 Comparison with Our Previous Results .............. 23
4.5 Discussions ....................................... 29
Chapter 5 Conclusions and Future Works .................... 33
References ................................................ 35
[1]L. A. Torre, F. Bray, R. L. Siegel, J. Ferlay, J. Lortet-Tieulent, and A. Jemal, "Global cancer statistics, 2012," CA: A Cancer Journal for Clinicians, vol. 65, no. 2, pp. 87-108, 2015.
[2]C. D. Mathers, D. M. Fat, and J. Boerma, The global burden of disease: 2004 update. World Health Organization, 2008.
[3]S. J. Otto, J. Fracheboud, A. L. Verbeek, R. Boer, J. C. I. Y. Reijerink-Verheij, J. D. M. Otten, M. J. M. Broeders, H. J. De Koning, and N. E. T. F. B. Can, "Mammography Screening and Breast Cancer Mortality-Response," (in English), Cancer Epidemiology Biomarkers & Prevention, vol. 21, no. 5, pp. 870-871, May 2012.
[4]M. Nothacker, V. Duda, M. Hahn, M. Warm, F. Degenhardt, H. Madjar, S. Weinbrenner, and U.-S. Albert, "Early detection of breast cancer: benefits and risks of supplemental breast ultrasound in asymptomatic women with mammographically dense breast tissue. A systematic review," Bmc Cancer, vol. 9, no. 1, p. 335, 2009.
[5]T. M. Kolb, J. Lichy, and J. H. Newhouse, "Occult cancer in women with dense breasts: detection with screening US--diagnostic yield and tumor characteristics," Radiology, vol. 207, no. 1, pp. 191-199, 1998.
[6]K. M. Kelly, J. Dean, W. S. Comulada, and S. J. Lee, "Breast cancer detection using automated whole breast ultrasound and mammography in radiographically dense breasts," (in English), European Radiology, vol. 20, no. 3, pp. 734-742, Mar 2010.
[7]W. A. Berg, Z. Zhang, D. Lehrer, R. A. Jong, E. D. Pisano, R. G. Barr, M. Bohm-Velez, M. C. Mahoney, W. P. Evans, L. H. Larsen, M. J. Morton, E. B. Mendelson, D. M. Farria, J. B. Cormack, H. S. Marques, A. Adams, N. M. Yeh, G. Gabrielli, and A. Investigators, "Detection of Breast Cancer With Addition of Annual Screening Ultrasound or a Single Screening MRI to Mammography in Women With Elevated Breast Cancer Risk," (in English), Jama-Journal of the American Medical Association, vol. 307, no. 13, pp. 1394-1404, Apr 4 2012.
[8]S. W. Chan, P. S. Cheung, S. Chan, S. S. Lau, T. T. Wong, M. Ma, A. Wong, and Y. C. Law, "Benefit of ultrasonography in the detection of clinically and mammographically occult breast cancer," World journal of surgery, vol. 32, no. 12, pp. 2593-2598, 2008.
[9]E. Fleury, J. Fleury, S. Piato, and D. Roveda Jr, "New elastographic classification of breast lesions during and after compression," Diagn Interv Radiol, vol. 15, no. 2, pp. 96-103, 2009.
[10]W. Rawat and Z. Wang, "Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review," (in eng), Neural Comput, pp. 1-98, Jun 09 2017.
[11]G. Litjens, T. Kooi, B. Ehteshami Bejnordi, A. Arindra Adiyoso Setio, F. Ciompi, M. Ghafoorian, J. A. W. M. van der Laak, B. van Ginneken, and C. I. Sanchez, "A Survey on Deep Learning in Medical Image Analysis," ArXiv e-prints, vol. 1702, p. arXiv:1702.05747, 2017.
[12]C. M. Lo, Y. P. Chen, Y. C. Chang, C. Lo, C. S. Huang, and R. F. Chang, "Computer-Aided Strain Evaluation for Acoustic Radiation Force Impulse Imaging of Breast Masses," (in English), Ultrasonic Imaging, vol. 36, no. 3, pp. 151-166, Jul 2014.
[13]C.-M. Lo, Y.-C. Chang, Y.-W. Yang, C.-S. Huang, and R.-F. Chang, "Quantitative breast mass classification based on the integration of B-mode features and strain features in elastography," Computers in Biology and Medicine, vol. 64, pp. 91-100, 2015.
[14]W. K. Moon, S.-C. Chang, J. M. Chang, N. Cho, C.-S. Huang, J.-W. Kuo, and R.-F. Chang, "Classification of Breast Tumors Using Elastographic and B-mode Features: Comparison of Automatic Selection of Representative Slice and Physician-Selected Slice of Images," Ultrasound in medicine & biology, 2013.
[15]A. Itoh, E. Ueno, E. Tohno, H. Kamma, H. Takahashi, T. Shiina, M. Yamakawa, and T. Matsumura, "Breast disease: clinical application of US elastography for diagnosis," Radiology, vol. 239, no. 2, pp. 341-50, May 2006.
[16]N. Nitta, M. Yamakawa, T. Shiina, E. Ueno, M. M. Doyley, and J. C. Bamber, "Tissue elasticity imaging based on combined autocorrelation method and 3-D tissue model," in 1998 IEEE Ultrasonics Symposium. Proceedings (Cat. No. 98CH36102), 1998, vol. 2, pp. 1447-1450 vol.2.
[17]K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition," ArXiv e-prints, vol. 1409, Accessed on: September 1, 2014Available: http://adsabs.harvard.edu/abs/2014arXiv1409.1556S
[18]S. Liu and W. Deng, "Very deep convolutional neural network based image classification using small training sample size," in 2015 3rd IAPR Asian Conference on Pattern Recognition (ACPR), 2015, pp. 730-734.
[19]J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, "Imagenet: A large-scale hierarchical image database," in Computer Vision and Pattern Recognition, 2009. CVPR 2009. IEEE Conference on, 2009, pp. 248-255: Ieee.
[20]A. L. Maas, A. Y. Hannun, and A. Y. Ng, "Rectifier nonlinearities improve neural network acoustic models," in Proc. ICML, 2013, vol. 30, no. 1.
[21]D. P. Kingma and J. Ba, "Adam: A Method for Stochastic Optimization," ArXiv e-prints, vol. 1412, Accessed on: December 1, 2014Available: http://adsabs.harvard.edu/abs/2014arXiv1412.6980K
[22]K. Chatfield, K. Simonyan, A. Vedaldi, and A. Zisserman, "Return of the Devil in the Details: Delving Deep into Convolutional Nets," ArXiv e-prints, vol. 1405, Accessed on: May 1, 2014Available: http://adsabs.harvard.edu/abs/2014arXiv1405.3531C
[23]A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
[24]K. Chatfield, K. Simonyan, A. Vedaldi, and A. Zisserman, "Return of the Devil in the Details: Delving Deep into Convolutional Nets," in ArXiv e-prints vol. 1405, ed, 2014.
[25]A. Krizhevsky, I. Sutskever, and G. E. Hinton, "ImageNet classification with deep convolutional neural networks," presented at the Proceedings of the 25th International Conference on Neural Information Processing Systems, Lake Tahoe, Nevada, 2012.
[26]Y. Min-Chun, M. Woo Kyung, Y. C. Wang, B. Min Sun, H. Chiun-Sheng, C. Jeon-Hor, and C. Ruey-Feng, "Robust Texture Analysis Using Multi-Resolution Gray-Scale Invariant Features for Breast Sonographic Tumor Diagnosis," (in eng), IEEE Trans Med Imaging, vol. 32, no. 12, pp. 2262-73, Dec 2013.
[27]W. C. Shen, R. F. Chang, and W. K. Moon, "Computer aided classification system for breast ultrasound based on breast imaging reporting and data system (BI-RADS)," (in English), Ultrasound in Medicine and Biology, vol. 33, no. 11, pp. 1688-1698, Nov 2007.
[28]R. M. Haralick, Shanmuga.K, and I. Dinstein, "Textural Features for Image Classification," (in English), Ieee Transactions on Systems Man and Cybernetics, vol. Smc3, no. 6, pp. 610-621, 1973.
[29]K. He, X. Zhang, S. Ren, and J. Sun, "Deep Residual Learning for Image Recognition," ArXiv e-prints, vol. 1512, p. arXiv:1512.03385, 2015.
[30]F. Chollet, "Xception: Deep Learning with Depthwise Separable Convolutions," ArXiv e-prints, vol. 1610, p. arXiv:1610.02357, 2016.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊