臺灣博碩士論文加值系統

全乳房自動超音波已被廣泛應用在乳房攝影，用以篩檢乳房病變。然而，檢視數百張的全乳房自動超音波影像十分耗時。因此，本研究提出一套基於三維卷積神經網路 (convolutional neural networks, CNN) 與優先候選聚合的電腦輔助偵測系統，輔助加速檢閱。首先，本系統採用有效的三維滑動視窗法掃描整個影像以提取三維的感興趣區域 (volumes of interest, VOI)。接著，三維 CNN 會估測每個 VOI 是腫瘤的機率，並且機率大於某閾值的 VOI 會留下成為候選腫瘤。因為候選腫瘤彼此可能重疊，本系統採用一套新穎的方法以聚合重疊的腫瘤。聚合時，腫瘤機率高者會賦予較高優先權，以避免過度聚合之問題。最後，上述步驟會針對不同目標腫瘤大小，重複執行數次。本研究有效利用目標腫瘤與 VOI 之間的相對大小關係，以優化各個步驟之效能。在偵測率 95% (184/194)、90% (175/194)、85% (165/194)、80% (155/194) 的情況下，本系統在每個全乳房自動超音波掃描平均分別產生 2.9、2.0、1.3 以及 0.8 個誤判區域。總結之，相較過去的方法，本研究設計的系統較為通用，執行速度較快，偵測效果也較佳。

Automated whole breast ultrasound (ABUS) has been widely used as a screening modality for examination of breast abnormalities. Reviewing hundreds of slices produced by ABUS, however, is time-consuming. Therefore, in this study, a fast and effective computer-aided detection (CADe) system based on 3-D convolutional neural networks (CNN) and prioritized candidate aggregation is proposed to accelerate this reviewing. Firstly, an efficient sliding window method is used to extract volumes of interest (VOIs). Then, each VOI is estimated the tumor likelihood with a 3-D CNN, and VOIs with higher estimated likelihood are selected as tumor candidates. Since the candidates may overlap each other, a novel scheme is designed to aggregate the overlapped candidates. During the aggregation, candidates are prioritized based on estimated tumor likelihood to alleviate over-aggregation issue. The relationship between the sizes of VOI and target tumor is optimally exploited to effectively perform each stage of our detection algorithm. On evaluation with a testing set of 192 tumors, our method achieved sensitivities of 95% (184/194), 90% (175/194), 85% (165/194), and 80% (155/194) with 2.9, 2.0, 1.3, and 0.8 FPs per pass, respectively. In summary, our method is more general and much faster than preliminary works, and demonstrates promising results.

口試委員審定書 i
致謝 ii
摘要 iii
Abstract iv
Table of Contents v
List of Figures vi
List of Tables vii
Chapter 1. Introduction 1
Chapter 2. Materials 4
Chapter 3. Methods 6
3.1 VOI Extraction with Sliding Window 8
3.2 Tumor Likelihood Estimation with 3-D CNN 9
3.2.1. Training 3-D CNN 11
3.3 Prioritized Candidate Aggregation 13
3.3.1. Alleviation of Over-aggregation 14
3.3.2. Multi-scale Aggregation 15
Chapter 4. Experimental Results and Discussion 16
4.1 Evaluation Methodology 16
4.2 Experimental Results 17
4.3 Discussion 23
Chapter 5. Conclusion 31
References 32

1.Siegel, R.L., K.D. Miller, and A. Jemal, Cancer statistics, 2016. CA Cancer J Clin, 2016. 66(1): p. 7-30.
2.Torre, L.A., et al., Global cancer statistics, 2012. CA Cancer J Clin, 2015. 65(2): p. 87-108.
3.Kolb, T.M., J. Lichy, and J.H. Newhouse, Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology, 2002. 225(1): p. 165-75.
4.Sickles, E.A., R.A. Filly, and P.W. Callen, Benign breast lesions: ultrasound detection and diagnosis. Radiology, 1984. 151(2): p. 467-70.
5.Roubidoux, M.A., et al., Bilateral breast cancer: early detection with mammography. Radiology, 1995. 196(2): p. 427-31.
6.Wilson, T.E., M.A. Helvie, and D.A. August, Breast cancer in the elderly patient: early detection with mammography. Radiology, 1994. 190(1): p. 203-7.
7.Chang, R.-F., et al., Whole breast computer-aided screening using free-hand ultrasound. International Congress Series, 2005. 1281(Complete): p. 1075-1080.
8.Ikedo, Y., et al., Development of a fully automatic scheme for detection of masses in whole breast ultrasound images. Med Phys, 2007. 34(11): p. 4378-88.
9.Moon, W.K., et al., Comparative study of density analysis using automated whole breast ultrasound and MRI. Medical Physics, 2011. 38(1): p. 382-389.
10.Chang, R.F., et al., Rapid image stitching and computer-aided detection for multipass automated breast ultrasound. Med Phys, 2010. 37(5): p. 2063-73.
11.Lo, C.M., et al., Multi-Dimensional Tumor Detection in Automated Whole Breast Ultrasound Using Topographic Watershed. IEEE Transactions on Medical Imaging, 2014. 33(7): p. 1503-1511.
12.Moon, W.K., et al., Computer-aided tumor detection based on multi-scale blob detection algorithm in automated breast ultrasound images. IEEE Trans Med Imaging, 2013. 32(7): p. 1191-200.
13.Tan, T., et al., Computer-aided detection of cancer in automated 3-D breast ultrasound. IEEE Trans Med Imaging, 2013. 32(9): p. 1698-706.
14.Lecun, Y., et al., Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998. 86(11): p. 2278-2324.
15.Krizhevsky, A., I. Sutskever, and G.E. Hinton, ImageNet classification with deep convolutional neural networks, in Proceedings of the 25th International Conference on Neural Information Processing Systems. 2012, Curran Associates Inc.: Lake Tahoe, Nevada. p. 1097-1105.
16.de Vos, B., et al., ConvNet-Based Localization of Anatomical Structures in 3D Medical Images. IEEE Transactions on Medical Imaging, 2017.
17.Day, W.H. and H. Edelsbrunner, Efficient algorithms for agglomerative hierarchical clustering methods. Journal of classification, 1984. 1(1): p. 7-24.
18.Sibson, R., SLINK: an optimally efficient algorithm for the single-link cluster method. The computer journal, 1973. 16(1): p. 30-34.
19.Simonyan, K. and A. Zisserman, Very Deep Convolutional Networks for Large-Scale Image Recognition. CoRR, 2014. abs/1409.1556.
20.He, K., et al., Deep Residual Learning for Image Recognition. CoRR, 2015. abs/1512.03385.
21.Srivastava, N., et al., Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res., 2014. 15(1): p. 1929-1958.
22.Nair, V. and G.E. Hinton. Rectified linear units improve restricted boltzmann machines. in Proceedings of the 27th international conference on machine learning (ICML-10). 2010.
23.Maas, A.L., A.Y. Hannun, and A.Y. Ng. Rectifier nonlinearities improve neural network acoustic models. in Proc. ICML. 2013.
24.Kingma, D. and J. Ba, Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
25.Chakraborty, D.P., Maximum likelihood analysis of free‐response receiver operating characteristic (FROC) data. Medical physics, 1989. 16(4): p. 561-568.
26.Chakraborty, D.P., et al., Digital and conventional chest imaging: a modified ROC study of observer performance using simulated nodules. Radiology, 1986. 158(1): p. 35-39.
27.Holm, S., A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics, 1979: p. 65-70.
28.Hochreiter, S. and J. Schmidhuber, Long short-term memory. Neural computation, 1997. 9(8): p. 1735-1780.
29.Girshick, R. Fast r-cnn. in Proceedings of the IEEE international conference on computer vision. 2015.
30.Ren, S., et al. Faster R-CNN: Towards real-time object detection with region proposal networks. in Advances in neural information processing systems. 2015.
31.Girshick, R., et al. Rich feature hierarchies for accurate object detection and semantic segmentation. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2014.
32.Uijlings, J.R., et al., Selective search for object recognition. International journal of computer vision, 2013. 104(2): p. 154-171.
33.Long, J., E. Shelhamer, and T. Darrell. Fully convolutional networks for semantic segmentation. in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015.