本論文提出一種以深度學習為基礎的自動且大規模的肝臟及其病變預測的 模型。以 MICCAI 2017 競賽的電腦斷層掃瞄圖做為我們的訓練資料，並藉由各 種前處理、數據增強及dropout 等方法增加我們的訓練資料以及解決深度學習常 遇見的過擬合問題。我們在肝臟預測的實驗結果得到一個不錯的結果，在預測階 段DICE Score 可以來到 91%，與其他參與者的分數不分上下，但在病變預測的 結果仍有改善的空間，與其他研究者有段小差距，原因可能為設備的不足，因為 設備的極限，導致我們在訓練時無法取得更多的特徵圖。在未來希望再提升設備 以及再改善我們的模型，讓病變預測的環節能有更大的進步。
53
參考文獻
[1] D. Ciresan, L. M. Gambardella, A. Giusti, J. Schmidhuber,” Deep neural networks segment neuronal membranes in electron microscopy images,” In: NIPS. pp.2852-2860, 2012.
[2] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan,V.
Vanhoucke, and A. Rabinovich,” Going deeper with convolutions,” in CoRR, abs/1409.4842, 2014.
[3] J. Ferlay, H.-R. Shin, F. Bray, D. Forman, C. Mathers, D. M. Parkin, Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127:
2893–917.
[4] A. Krizhevsky, I. Sutskever, G.E. Hinton, “Imagenet classification with deep convolutional neural networks,” In Advances in Neural Information Processing Systems 25, pages 1106–1114, 2012.
[5] K. He, X. Zhang, S. Ren, J. Sun, “Deep Residual Learning for Image Recognition,”
In:CVPR, arXiv preprint arXiv:1512.03385, 2015.
[6] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, A. L. Yuille, “Semantic image segmentation with deep convolutional nets and fully connected crfs,” In:
ICLR, 2015.
[7] G. Li, X. Chen, F. Shi, W. Zhu, J. Tian, D. Xiang, “Automatic liver segmentation based on shape constraints and deformable graph cut in ct images,” IEEE Trans.
Image Process, vol.24, pp. 5315–5329, 2015.
[8] J. Long, E. Shelhamer, T. Darrell, “Fully convolutional networks for semantic segmentation,” CoRR, abs/1411.4038, 2014.
54
[9] M. Havaei, A. Davy, D. Warde-Farley, A. Biard, A. Courville, Y. Bengio, C. Pal, P.-M. Jodoin, H. Larochelle, “Brain Tumor Segmentation with Deep Neural Networks,” CVPR, arXiv preprint arXiv:1505.03540, 2015.
[10] P. Ferdinand Christ et al., “Automatic Liver and Tumor Segmentation of CT and MRI Volumes using Cascaded Fully Convolutional Neural Networks,” CVPR, arXiv preprint arXiv:1702.05970, 2017.
[11] O. Ronneberger, P. Fischer, T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in: MICCAI, Vol. 9351, pp. 234-241, 2015.
[12] M. Seyedhosseini, M. Sajjadi, T. Tasdizen, “Image segmentation with cascaded hierarchical models and logistic disjunctive normal networks,” In: IEEE International Conference on Computer Vision (ICCV), pp. 2168–2175, 2013.
[13] K. Simonyan, A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” CoRR, abs/1409.1556, 2014.
[14] S. Zheng, S. Jayasumana, B. Romera-Paredes, V. Vineet, Z. Su, D. Du, C. Huang, P. Torr, “Conditional Random fields as recurrent neural networks,” in ICCV, 2015.
[15] T. Heimann, et al., ”Comparison and evaluation of methods for liver segmentation from ct datasets,” IEEE Transactions on Medical Imaging 28, vol 8., pp. 1251-1265. doi:10.1109/TMI.2009.2013851, 2009.
[16] R. Vivanti, L. Joskowicz, A. Ephrat, J. Sosna, “Automatic liver tumor segmentation in follow-up CT studies using Convolutional Neural Networks,” in MICCAI 2015, Munich, Germany, 2015.
[17] Y. Le Cun, B. Boser, J.S. Denker, D. Henderson, R.E. Howard, W. Hubbard, L.D.
Jackel, et al., “Handwritten digit recognition with a back-propagation network. In Advances in neural information processing systems,” 1990.
55
[18] ] X. Deng and G. Du, “Editorial: 3-D segmentation in the clinic: A grand challenge II—Liver tumor segmentation,” in Proc. MICCAI Workshop on 3-D Segmentation in the Clinic, 2008
[19] Y. Hame and M. Pollari, “Semi-automatic liver tumor segmentation with hidden markov measure field model and non-parametric distribution estimation,” Medical Image Analysis 16, vol.1, pp.140-149, 2012.
[20] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner,” Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no.11, pp. 2278–2324, 1998.
[21] D.O. Hebb,”The Organization of Behavior,” New York: Wiley & Sons, 1949.
[22] J.C. Duchi, E. Hazan, and Y. Singer, “Adaptive subgradient methods for online learning and stochastic optimization,” Journal of Machine Learning Research, 2011.
[23] Kingma, Diederik P. and Ba, Jimmy,”Adam: A Method for Stochastic Optimization,” arXiv:1412.6980 [cs.LG], 2014.
[24] N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov,”
Dropout: A simple way to prevent neural networks from overfitting,” JMLR, vol.15, pp.1929–1958, 2014.
[25] T. Acharya and A.K. Ray, “Image Processing Principles and Applications,” New Jersey: John Wiley & Sons, Inc, 2005.
[26] G. Chlebus, H. Meine, J. H. Moltz, and A. Schenk, “Neural network-based automatic liver tumor segmentation with random forest-based candidate filtering,” arXiv preprint arXiv:1706.00842, 2017
[27] X. Han, “Automatic liver lesion segmentation using a deep convolutional neural network method,” arXiv preprint arXiv:1704.07239.
[28] https://competitions.codalab.org/
56
附錄
開發環境 Tensorflow-keras
程式語言 Python3.5
本 論 文 程 式 是 採 用 tensorflow 為 後 置 運 算 ， 因 此 一 開 始 先 至
https://www.tensorflow.org/install/install_window
此 官 網 ， 按 照 其 步 驟 安 裝 tensorflow，再來我們是以 keras 來建構本論文中使用的網路架構，因此 python 需 要再擴增keras 等的 API，以及各種圖片處理的函式庫。下面分為兩個程式，第 一個主要是處理各式前處理，包括數據增強以及直方圖均化等等。第二個程式為 處理訓練以及預測。程式ㄧ：
以上為此程式所要用到的函式庫。
57
此function 為尋找此切片圖是否有肝臟或者病變，以及算出其的個數。
58
以上為數據增強的程式碼，可以依照需求調整各項的參數。
59
以上為直方圖均化的function。
60
以上為我們創造訓練資料的源碼。
61
此為之後要做測試的測試數據。
62
程式二：
以上為此程式用到的函式庫。
以上為本文使用的DICE Score 的程式碼。
63
此function 為本文所使用的網路架構的前半段，為一般的卷積神經網路架構。
接續上面的程式碼，此為網路結構的下半段，這裡開始使用Upsampling 的方法。
64
本function 主要工作為訓練以及預測，並把最後的結果儲存起來。