在本論文中，我們提出了一個結合肺部影像分割網路和一個基於DoubleUnet的肺結節分割網路，並採用了使用不同損失函數的深度監督機制。肺結節影像分割網路是一種應用於肺部CT影像的深度學習模型，主要目標是以pixel為單位標出肺結節。我們提出的模型使用肺部影像分割網路的輸出去限制肺結節網路尋找的影像視野，再藉DoubleUnet二段式網路的特性結合兩種不同loss，讓模型能在不同階段擁有不同的特性，讓網路得以在適當的範圍內精準地完成任務。最終於公開資料集LIDC-IDRI取得62.26 %的IoU分數，並於私有資料集TMUH取得68.32%的IoU分數，相較原始DoubleUnet分別提高了6.71%與6.2%。最後我們也比較並分析了模型在私有與公開資料集間的表現差異，我們也基於這些結果，對未來醫學影像分割任務提出一些訓練過程的建議。

In this thesis, a deep learning model that combines a lung segmentation network and a DoubleUnet based nodule segmentation network is proposed and a deep supervision mechanism with different loss functions is employed. The lung nodule segmentation network is a deep learning model applied to lung CT images. Its main target is to mark out the nodule pixel-by-pixel. The proposed model uses the output of a lung segmentation network to limit the field of view of the nodule segmentation network, and with the two-stage characteristics of DoubleUnet, our deep supervision with different losses lets the model have different learning characteristics on different stages. With these modifications, the model can accurately finish its job within appropriate regions. The proposed model reaches 68.32% of IoU score on the TMUH private dataset, and 62.26% of IoU score on the LIDC-IDRI public dataset, which has an improvement of 6.2% and 6.71%, respectively, compared with the results of using original DoubleUnet. In addition, we compared and analyzed the performance differences between the public and the private dataset. Based on these results, some suggestions on the training process of medical image segmentation tasks are provided for future research on heterogeneous databases.

[1] O. Ronneberger, P. Fischer and T. Brox, "U-Net: Convolutional networks for biomedical image segmentation," in MICCAI, 2015.
[2] D. Jha, M. Riegler, D. Johansen, P. Halvorsen and H. D. Johansen, "DoubleU-Net: a deep convolutional neural network for medical image segmentation," 2020 IEEE 33rd International Symposium on Computer-Based Medical Systems (CBMS), pp. 558-564, 2020.
[3] A. Mansoor, U. Bagci, B. Foster, Z. Xu, G. Papadakis, L. Folio, J. Udupa and D. Mollura, "Segmentation and image analysis of abnormal lungs at CT: current approaches, challenges, and future trends," Radiographics : a review publication of the Radiological Society of North America, Inc, vol. 35 4, pp. 1056-76, 2015.
[4] E. Shelhamer, J. Long and T. Darrell, "Fully convolutional networks for semantic segmentation," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 39, pp. 640-651, 2017.
[5] K. He, G. Gkioxari, P. Dollár and R. B. Girshick, "Mask R-CNN," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 42, pp. 386-397, 2020.
[6] L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy and A. Yuille, "DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, pp. 834-848, 2018.
[7] X.-n. Xiao, S. Lian, Z. Luo and S. Li, "Weighted Res-UNet for high-quality retina vessel segmentation," 2018 9th International Conference on InformationTechnology in Medicine and Education (ITME), pp. 327-331, 2018.
[8] K. He, X. Zhang, S. Ren and J. Sun, "Deep residual learning for image recognition," 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770-778, 2016.
[9] S. Guan, A. A. Khan, S. Sikdar and P. Chitnis, "Fully dense UNet for 2-D sparse photoacoustic tomography artifact removal," IEEE Journal of Biomedical and Health Informatics, vol. 24, pp. 568-576, 2020.
[10] G. Pleiss, D. Chen, G. Huang, T. Li, L. V. D. Maaten and K. Q. Weinberger, "Memory-Efficient implementation of DenseNets," ArXiv, vol. abs/1707.06990, 2017.
[11] S. R. Ravichandran, B. Nataraj, S. Huang, Z. Qin, Z. Lu, A. Katsuki, W. Huang and Z. Zeng, "3D Inception U-Net for aorta segmentation using computed tomography cardiac angiography," 2019 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI), pp. 1-4, 2019.
[12] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. E. Reed, D. Anguelov, D. Erhan, V. Vanhoucke and A. Rabinovich, "Going deeper with convolutions," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1-9, 2015.
[13] Z. Zhou, M. R. Siddiquee, N. Tajbakhsh and J. Liang, "UNet++: a nested U-Net architecture for medical image segmentation," 4th Deep Learning in Medical Image Analysis (DLMIA) Workshop, vol. 11045, pp. 3-11, 2018.
[14] H. Liu, X. Shen, F. Shang and F. Wang, "CU-Net: Cascaded U-Net with loss weighted sampling for brain tumor segmentation," ArXiv, vol. abs/1907.07677, 2019.
[15] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," CoRR, vol. abs/1409.1556, 2015.
[16] J. Hu, L. Shen and G. Sun, "Squeeze-and-Excitation networks," 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7132-7141, 2018.
[17] Ö. Çiçek, A. Abdulkadir, S. Lienkamp, T. Brox and O. Ronneberger, "3D U-Net: learning dense volumetric segmentation from sparse annotation," ArXiv, vol. abs/1606.06650, 2016.
[18] X. Li, H. Chen, X. Qi, Q. Dou, C.-W. Fu and P. Heng, "H-DenseUNet: hybrid densely connected UNet for liver and tumor segmentation From CT volumes," IEEE Transactions on Medical Imaging, vol. 37, pp. 2663-2674, 2018.
[19] F. Milletari, N. Navab and S.-A. Ahmadi, "V-Net: fully convolutional neural networks for volumetric medical image segmentation," 2016 Fourth International Conference on 3D Vision (3DV), pp. 565-571, 2016.
[20] Z. Zhou, V. Sodha, M. R. Siddiquee, R. Feng, N. Tajbakhsh, M. Gotway and J. Liang, "Models Genesis: generic autodidactic models for 3D medical image analysis," MICCAI, vol. 11767, pp. 384-393, 2019.
[21] F. Haghighi, M. Taher, Z. Zhou, M. Gotway and J. Liang, "Learning semantics-enriched representation via self-discovery, self-classification, and self-restoration," in MICCAI, 2020.
[22] E. Rendon-Gonzalez and V. Ponomaryov, "Automatic lung nodule segmentation and classification in CT images based on SVM," 2016 9th International Kharkiv Symposium on Physics and Engineering of Microwaves, Millimeter and Submillimeter Waves (MSMW), pp. 1-4, 2016.
[23] Y. Zhu and C. Huang, "An adaptive histogram equalization algorithm on the image gray level mapping," Physics Procedia, vol. 25, pp. 601-608, 2012.
[24] J. Yang, H. Veeraraghavan, S. Armato, K. Farahani, J. Kirby, J. Kalpath™ Kramer, W. V. v. Elmpt, A. Dekker, X. Han, X. Feng, P. Aljabar, B. Oliveira, B. v. d. Heyden,L. Zamdborg, D. Lam, M. Gooding and G. Sharp, "Autosegmentation for thoracic radiation treatment planning: a grand challenge at AAPM 2017," Medical Physics, vol. 45, p. 4568 4581, 2018.
[25] G. Huang, Z. Liu and K. Q. Weinberger, "Densely connected convolutional networks," 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2261-2269, 2017.
[26] C.-Y. Lee, S. Xie, P. W. Gallagher, Z. Zhang and Z. Tu, "Deeply-Supervised nets," ArXiv, vol. abs/1409.5185, 2015.
[27] Y. Zhang and A. C. S. Chung, "Deep supervision with additional labels for retinal vessel segmentation task," in MICCAI, 2018.
[28] S. Armato, G. McLennan, L. Bidaut, M. McNitt-Gray, C. Meyer, A. Reeves, B. Zhao, D. Aberle, C. Henschke, E. Hoffman, E. Kazerooni, H. MacMahon, E. J. R. V. Beeke, D. Yankelevitz, A. Biancardi, P. Bland, M. S. Brown, R. Engelmann, G. Laderach, D. Max, R. C. Pais, D. Qing, R. Roberts, A. R. Smith, A. Starkey, P. Batrah, P. Caligiuri, A. O. Farooqi, G. Gladish, C. Jude, R. Munden, I. Petkovska, L. Quint, L. Schwartz, B. Sundaram, L. Dodd, C. Fenimore, D. Gur, N. Petrick, J. Freymann, J. Kirby, B. Hughes, A. V. Casteele, S. Gupte, M. Sallamm, M. Heath, M. Kuhn, E. Dharaiya, R. Burns, D. Fryd, M. Salganicoff, V. Anand, U. Shreter, S. Vastagh and B. Croft, "The lung image database consortium (LIDC) and image database resource initiative (IDRI): a completed reference database of lung nodules on CT scans," Medical physics, vol. 38 2, pp. 915-31, 2011.
[29] D. P. Kingma and J. Ba, "Adam: A Method for Stochastic Optimization," CoRR, vol. abs/1412.6980, 2015.
[30] "Papers with code - the latest in machine learning." https://paperswithcode.com/ (accessed 02, August, 2021).