肺癌是台灣的主要癌症死因之一，具有高發病率和高致死率。低劑量電腦斷層掃描(Low-Dose Computed Tomography, LDCT, LDCT)是目前證實最有效的肺癌早期篩檢工具。而衛福部自2022年起推行肺癌早期偵測計畫，提供高風險族群每兩年一次的低劑量電腦斷層掃描檢查。雖有效果，但大量新增影像增加了醫師的判讀負擔。為了減輕這一負擔，本論文提出利用深度學習模型進行電腦輔助診斷，但因單一醫院的資料量有限，難以訓練出準確且穩定的模型。此外，由於醫療數據共享面臨法律與隱私保護的挑戰，傳統的集中式訓練方法難以實施。為此，本研究引入聯邦學習(Federated Learning, FL)作為解決方案，該方法允許不同醫院在不共享敏感資料的情況下進行協同模型訓練，從而保護患者隱私並克服資料稀缺問題。同時，為解決聯邦學習中依賴完整標記資料的挑戰，本論文採用了基於老師學生架構的半監督學習(Semi-Supervised Learning)技術，通過生成偽標籤使未標記資料參與訓練，進一步提高模型的準確性與穩定性。本論文最終構建了一個跨醫院的3D肺結節偵測系統，該系統結合了國立成功大學附設醫院、嘉義基督教醫院與台南市立醫院三家醫療機構的資源，利用聯邦學習和半監督學習技術，在保護患者隱私的前提下有效提升模型性能，減輕醫師的診斷負擔。實驗結果顯示，本論文所提出的方法在成大醫院的測試資料集上達到了88.5%的召回率，當僅針對大於4mm的肺結節時，召回率更提升至89.6%。這表明，通過結合聯邦學習與半監督學習，系統在少量標記資料的情況下仍能有效利用未標記資料進行高效訓練，成功地應用於肺結節偵測，對於提升診斷效率和減輕醫師工作負擔具有顯著意義。

Lung cancer is a leading cause of cancer-related deaths in Taiwan, with high incidence and mortality rates. Low-Dose Computed Tomography (LDCT) is the most effective tool for early detection of lung cancer. Since 2022, the Ministry of Health and Welfare of Taiwan has provided LDCT screenings every two years to high-risk groups. While this initiative has proven effective, the increased volume of images has significantly added to physicians' diagnostic workload. To address this, this thesis proposes using deep learning models for computer-aided diagnosis. However, limited data from single hospitals and legal challenges in data sharing complicate traditional centralized training methods. To overcome these issues, this study introduces Federated Learning (FL), which enables collaborative model training without sharing sensitive data, thus protecting patient privacy and addressing data scarcity.Additionally, the paper employs a semi-supervised learning (SSL) approach, using a teacher-student framework to generate pseudo-labels and utilize unlabeled data, enhancing model accuracy and stability. The developed cross-hospital 3D lung nodule detection system, integrating resources from National Cheng Kung University Hospital, Chiayi Christian Hospital, and Tainan City Hospital, effectively combines FL and SSL. This approach improves model performance while safeguarding patient privacy and reduces physicians' diagnostic burden. Experimental results show an 88.5% recall rate on the test dataset, which increases to 89.6% for nodules larger than 4 mm, demonstrating the method's effectiveness in enhancing diagnostic efficiency and reducing physician workload.

[1] K. Bonawitz, H. Eichner, W. Grieskamp, D. Huba, A. Ingerman, V. Ivanov, C. Kiddon, J. Koneˇcn´y, S. Mazzocchi, H.B. McMahan, T.V. Overveldt, D. Petrou, D. Ramage and J. Roselander, "Towards Federated Learning at Scale: System Design," In Proceedings of Machine Learning and Systems, pp. 374-388, 2019.
[2] L. Daliang and J. Wang, "Fedmd: Heterogenous Federated Learning via Model Distillation," ArXiv preprint arXiv:1910.03581, 2019.
[3] M. Dolejˇs´ and J. Kybic, "Automatic Two-step Detection of Pulmonary Nodules," In Medical Imaging 2007: Computer-Aided Diagnosis, vol. 6514, pp. 1093-1104, 2007.
[4] F. Garcea, A. Serra, F. Lamberti and L. Morra, "Data Augmentation for Medical Imaging: A Systematic Literature Review," Computers in Biology and Medicine, 152, pp. 106391, 2023.
[5] J. Hu, L. Shen and G. Sun, "Squeeze-and-excitation Networks," In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 7132-7141, 2018.
[6] W. Jeong, J. Yoon, E. Yang and S.J. Hwang, "Federated Semi-Supervised Learning with Inter-Client Consistency & Disjoint Learning," In International Conference on Learning Representations (ICLR), 2021.
[7] D.P. Kingma and J.L. Ba, "Adam: A Method for Stochastic Optimization," ArXiv preprint arXiv:1412.6980, 2014.
[8] H. Kaiming, X. Zhang, S. Ren and J. Sun, "Deep Residual Learning for Image Recognition," In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770-778, 2016.
[9] Y.C. Liu, C.Y. Ma and Z. Kira1, "Unbiased Teacher v2: Semi-supervised Object Detection for Anchor-free and Anchor-based Detectors," In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 9819-9828, 2022.
[10] T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar and V. Smith, "Federated Optimization in Heterogeneous Networks," Proceedings of Machine Learning and Systems, 2, pp. 429-450, 2020.89
[11] X. Luo, T. Song, G. Wang, J. Chen, Y. Chen, K. Li, D.N. Metaxas and S. Zhang, "SCPM-Net: An Anchor-free 3D Lung Nodule Detection Network Using Sphere Representation and Center Points Matching," In Medical Image Analysis, 75, p.102287, 2022.
[12] H.B. McMahan, E. Moore, D. Ramage, S. Hampson, B.A.Y. Arcas, "Communicationefficient Learning of Deep Networks from Decentralized Data," In Artificial Intelligence and Statistics, pp. 1273-1282, 2017.
[13] J. Mei, M.M. Cheng, G. Xu, L.R. Wan and H. Zhang, “SANet: A Slice-Aware Network for Pulmonary Nodule Detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 44, pp. 4374-4387, 2021.
[14] S.J. Reddi, Z. Charles , M. Zaheer, Z. Garrett, K. Rush, J. Konecný, S. Kumar and H. Brendan McMahan, "Adaptive Federated Optimization," In International Conference on Learning Representations (ICLR), 2021.
[15] T. Song, J. Chen, X. Luo, Y. Huang, X. Liu, N. Huang, Y. Chen, Z. Ye, H. Sheng, S. Zhang and G. Wang, "CPM-Net: A 3D Center-points Matching Network for Pulmonary Nodule Detection in CT Scans, " In International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), pp. 550-559, 2020.
[16] K. Sohn, Z. Zhang, C. Li, H. Zhang, C. Lee and T. Pfister, "A Simple Semi-supervised Learning Framework for Object Detection," ArXiv preprint arXiv:2005.04757, 2020.
[17] A. Shrivastava, A. Gupta and R. Girshick, “Training Region-Based Object Detectors with Online Hard Example Mining,” In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 761-769, 2016.
[18] K. Sohn, D. Berthelot, C. Li, Z. Zhang, N. Carlini, E.D. Cubuk, A. Kurakin, H. Zhang and C. Raffel, "Fixmatch: Simplifying Semi-Supervised Learning with Consistency and Confidence," In Neural Information Processing Systems (NeurIPS), 33, pp. 596-608, 2020.
[19] D. Wang, Z. Yuan, K. Zhang and L. Wang, "Focalmix: Semi-supervised Learning for 3D Medical Image Detection," In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3951-3960, 2020.
[20] A. Xu, W. Li, P. Guo, D. Yang, H. Roth, A. Hatamizadeh, C. Zhao, D. Xu, H. Huang, and Z. Xu, "Closing the Generalization Gap of Cross-silo Federated Medical Image 90 Segmentation," Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 20866-20875, 2022.
[21] M. Xu, Z. Zhang, H. Hu, J. Wang, L. Wang, F. Wei and Z. Liu, "End-to-End SemiSupervised Object Detection with Soft Teacher, " In Proceedings of the IEEE/CVF International Conference on Computer Vision (CVPR), pp. 3060-3069, 2021.
[22] Z. Yang, Z. Li, X. Jiang, Y. Gong, Z. Yuan, D. Zhao and C. Yuan, “Focal and Global Knowledge Distillation for Detectors,” In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4643-4652, 2022.
[23] Z. Zhang, Y. Yang, Z. Yao, Y. Yan, J. E. Gonzalez, K. Ramchandran and M.W. Mahoney, "Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models," In 2021 IEEE International Conference on Big Data (Big Data), pp. 1214-1225, 2021.
[24] 阮青龍, "使用改進的 Transformer 和時空對應網絡之交互式分割技術來標記 3D 肺結節," 成 功 大 學 碩 圖 書 館, 2023. [Online] Availabel: https://thesis.lib.ncku.edu.tw/thesis/detail/14416de677104d1d36ffbe423c6d9e0f/.
[25] "Open Federated Learning Documentation," Intel OpenFL, 2024. [Online]. Available: https://openfl.readthedocs.io/en/latest/.
[26] "Intel Distribution of Openvino Toolkit," Intel Openvino, 2024. [Online]. Available: https://www.intel.com/content/www/us/en/developer/tools/openvinotoolkit/overview.html.