基於聯邦式學習的U-Net肺結節分割性能優化研究

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：104

、訪客IP：18.119.163.95

姓名

許子麒(Tzu-Chi Hsu) 查詢紙本館藏

畢業系所

通訊工程學系

論文名稱

基於聯邦式學習的U-Net肺結節分割性能優化研究
(Research on optimization of U-Net pulmonary nodule segmentataion performance based on federated learning)

相關論文

★ UHF頻段RFID彈藥管理系統之設計、實作與評估	★ 移動物偵測與追蹤之IP Camera系統
★ SDN自適應性自動化網路安全之研究	★ Wi-Fi Direct Service 應用於IoT
★ 射頻前端電路應用於載波聚合長期演進技術	★ 3C無線充電裝置運用在車載系統所產生之EMI輻射
★ 基於LoRa技術的物聯網前端防盜警示感測裝置實作與評估	★ DOCSIS 3.1 效能研究與下行通道干擾阻隔之設計
★ 藍芽無線光學投影翻譯筆	★ 手持裝置應用於MIMO ( 8x8 ) Wi-Fi系統之設計
★ 基於無伺服器運算之智慧農業雲端系統設計與研究	★ 嵌入式系統實現電梯物聯網
★ 在802.11 Ad-Hoc網路中基於速率考量之路由協定設計	★ 合作博弈與灰色模糊方法改善無線網路之性能
★ 採用拍賣策略之動態分散式方法於減少叢集小型基地台間干擾之研究	★ 在LTE-A下聚合未授權頻譜及動態分配資源以優化系統效能

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2026-7-10以後開放)

摘要(中)

近來，隨著深度神經網路技術的快速發展，其在醫療影像領域的應用也日益增多，其中肺結節分割模型訓練就是其中之一，但礙於醫療影像牽涉到個人隱私、合法性，無法彼此共享交流，偏鄉地區醫院的數據量較小，可能導致模型性能在訓練時無法達到最佳化，因此在這樣的前提下採用聯邦學習架構結合本地模型做訓練，會是最適合的選擇。
聯邦學習是一種新穎的機器學習方法，可以達到實現分散式學習的同時，也維護資料安全性。聯邦學習訓練中，將由伺服器端發送初始化模型給各參與聯邦的客戶端做本地訓練，且各個客戶端使用獨立的本地數據，彼此不共享隱私數據，僅藉由回傳模型訓練權重至伺服器端聚合，更新後的模型權重再回傳給客戶端做訓練，使模型能學習不同數據的多樣性，來提高整體的性能及可靠性。
本篇論文採用Flower作為模擬環境，並假設兩間不同地理位置的醫院，彼此數據分佈不均，藉由聯邦學習架構所帶來的數據多樣性，來優化最終分割的準確度。

摘要(英)

Recent advancements in deep neural network technologies have significantly increased their applications in medical imaging. Nonetheless, the sensitive nature of medical data and legal constraints prevent data sharing, particularly in rural areas where hospitals have limited data availability. This limitation can hinder the optimization of model training. Under these circumstances, federated learning provides an optimal solution by enabling local model training without data exchange, thereby maintaining data privacy.
Federated learning is a novel machine learning method that facilitates distributed learning while maintaining data security. In this process, a server sends an initial model to federated clients for local training. Each client uses their independent data without sharing private information. They then return their model parameters to the server for aggregation. The updated parameters are redistributed to the clients for further training, enabling the model to learn from diverse data, thus enhancing overall performance and reliability.
The paper adopts Flower as the simulation environment and assumes two hospitals in different geographical locations, with unevenly distributed data between them. By leveraging the data diversity brought by the federated learning framework, the aim is to optimize the final segmentation accuracy .

關鍵字(中)

★ U-Net
★ 深度學習
★ 聯邦式學習
★ 醫學影像
★ 數據不足

關鍵字(英)

★ U-Net
★ Deep Learning
★ Federated Learning
★ Medical Imaging
★ Data Scarcity

論文目次

目錄
摘要 i
Abstract ii
誌謝 iii
目錄 iv
圖目錄 vii
表目錄 xi
第一章序論 1
1-1 前言 1
1-2 研究動機 2
1-3 論文架構 3
第二章相關研究背景 4
2-1 卷積神經網路 4
2-1-1 卷積層 5
2-1-2 池化層 6
2-1-3 全連接層 7
2-1-4 Sigmoid激活函數 8
2-1-5 ReLU激活函數 10
2-2 語義分割網路 11
2-2-1 U-Net 12
2-2-2 收縮路徑 13
2-2-3 擴張路徑 14
2-2-4 反卷積 15
2-3 模型擬合 16
2-4 聯邦式學習 17
2-4-1 橫向聯邦學習 18
2-4-2 縱向聯邦學習 19
2-4-3 聯邦平均演算法 19
31 第三章數據預處理與流程 21
3-1 數據集 21
3-1-1 電腦斷層掃描影像 ( CT ) 22
3-2 預處理 23
3-2-1 HU值轉換 23
3-2-2 CLAHE直方圖均衡化 24
3-2-3 標準化 26
3-2-4 K-means聚類分割 27
3-2-5 二值化 29
3-2-6 肺部去噪填滿 30
3-3 數據預處理流程 32
3-4 數據分割 33
第四章模擬環境架構與結果 37
4-1 模擬環境 37
4-2 模擬方法 38
4-3 模型評估標準 42
4-4 模擬結果與分析 43
4-4-1 病例7:3本地端訓練結果 43
4-4-2 病例7:3聯邦訓練結果與比較 45
4-4-3 病例9:1本地訓練結果 50
4-4-4 病例9:1聯邦訓練結果與比較 51
第五章研究結果與討論 55
5-1 研究結果 55
5-2 限制性 56
5-3 結論 57
參考文獻 58

參考文獻

[1] Y Liu, B Bai, H.C. Chen, P Liu, H. M. Feng, “Cervical Image Segmentation using U-Net Model”, 2019 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS),2020.
[2] U. Aakanksha, K. Vijay Chandra , “Corneal Image Segmentation Using Attention U-Net”, 2024 Fourth International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT),2024.
[3] A Raj, V Sharma, A. K. Shanu, “Comparative Analysis Of Security And Privacy Technique For Federated Learning In IOT Based Devices”, 2022 3rd International Conference on Computation, Automation and Knowledge Management (ICCAKM),2022.
[4] Z Li, S He, P Chaturvedi, T.H. Hoang, M Ryu, E.A. Huerta, V Kindratenko, J Fuhrman, M Giger, R Chard, K Kim, R Madduri “APPFLx: Providing Privacy-Preserving Cross-Silo Federated Learning as a Service”, 2023 IEEE 19th International Conference on e-Science (e-Science),2023.
[5] K. L. D. Viet, K. L. Ha; T. N. Quoc, V.T. Hoang, “MRI Brain Tumor Classification based on Federated Deep Learning”, 2023 Zooming Innovation in Consumer Technologies Conference (ZINC),2023.
[6] W Lai and Q Yan, “Federated Learning for Detecting COVID-19 in Chest CT Images: A Lightweight Federated Learning Approach”, 2022 4th International Conference on Frontiers Technology of Information and Computer (ICFTIC),2023.

[7] K. S. Chethan, S. Vishwanath, R. V. Patil, K. A. Vijetha, “Segementation and Prediction from CT Images for Detecting Lung Cance”, 2020 11th International Conference on Computing, Communication and Networking Technologies (ICCCNT),2020.
[8] A Krizhevsky, I Sutskever, G Hinton, “ImageNet Classification with Deep Convolutional Neural Networks”, Communications of the ACM , Vol.60, Issue 6, pp.84-90,2012.
[9] N. Devi, B. Borah, “Cascaded pooling for Convolutional Neural Networks”, 2018 Fourteenth International Conference on Information Processing (ICINPRO), 2018.
[10] Khan, A., Sohail, A., Zahoora, U., & Qureshi, A. S. (2020). “A survey of the recent architectures of deep convolutional neural networks”, Artifical Intelligence Review, 53(8), 5455-5516.
[11] C. Zhao, B. Jeon, “Compact Representation of Light Field Data for Refocusing and Focal Stack Reconstruction Using Depth Adaptive Multi-CNN”, IEEE Transactions on Computational Imaging, vol.10, pp.170-180, 2024.
[12] A. Bagrodia, A. Shetty U., Shruthi M.L.J., "Pathological Lung Segmentation in Chest X-Ray Images using Modified U-Net-Based Architecture", 2023 IEEE 20th India Council International Conference (INDICON), pp.452-456, 2023.
[13] E. Shelhamer, J. Long, T. Darrell, “Fully Convolutional Networks for Semantic Segmentation”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 39, Issue. 4, pp.640-651,2016.
[14] 網路資料 on line resources:語義分割。取自 https://medium.com/ching-i/%E5%BD%B1%E5%83%8F%E5%88%86%E5%89%B2-image-segmentation-%E8%AA%9E%E7%BE%A9%E5%88%86%E5%89%B2-semantic-segmentation-1-53a1dde9ed92.
[15] O. Ronneberger, P. Fischer, T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation”, Computer vision and pattern Recognition(cs.cv),2015, https://doi.org/10.48550/arXiv.1505.04597.
[16] W. Liangsheng, L. Chentong, L. Xingjian, “ Deep Neural Networks for Semantic Segmentation of Lung Nodules”, 2021 40th Chinese Control Conference (CCC), July 2021.
[17] 網路資料on line resources: Overfitting and Underfitting。取自https://ithelp.ithome.com.tw/articles/10305467
[18] A. Ray, H. Ray, “Study of Overfitting through Activation Functions as a Hyper-parameter for Image Clothing Classification using Neural Network”, 2021 12th International Conference on Computing Communication and Networking Technologies (ICCCNT), 2021.
[19] P. Thanapol, K. Lavangnananda, P. Bouvry, F. Pinel, F. Leprévost, “Reducing Overfitting and Improving Generalization in Training Convolutional Neural Network (CNN) under Limited Sample Sizes in Image Recognition”, 2020 - 5th International Conference on Information Technology (InCIT), 2020.
[20] B. McMahan, D. Ramage, “Federated Learning: Collaborative Machine Learning without Centralized Training Data”, 2017, 取自https://research.google/blog/federated-learning-collaborative-machine-learning-without-centralized-training-data/
[21] W. Sing, H. Chen, Z. Qiu, L. Luo, “A Federated Learning Scheme Based on Lightweight Differential Privacy”, 2023 IEEE International Conference on Big Data (BigData), 2023.
[22] R. Zeng, B. Mi, D. Huang, “A Federated Learning Framework Based on CSP Homomorphic Encryption”, 2023 IEEE 12th Data Driven Control and Learning Systems Conference (DDCLS), 2023.
[23] Q. Yang, Y. Liu, T. Chen, Y. Tong, “Federated Machine Learning: Concept and Applications”, ACM Transactions on Intelligent Systems and Technology (TIST), Vol 10, Issue 2, Article No. 12, 2019.
[24] B. McMahan, E. Moore, D. Ramage, B. A. Y. Arcas, “Federated learning of deep networks using model averaging”, arXiv preprint arXiv:1602.05629 (2016).
[25] A. Korkmaz, A. Alhonainy, P. Rao, “An Evaluation of Federated Learning Techniques for Secure and Privacy-Preserving Machine Learning on Medical Datasets”, 2022 IEEE Applied Imagery Pattern Recognition Workshop (AIPR), 2023.
[26] D. Zhu, J. Chen, X. Zhou, W. Shang, A. E. Hassan, J. Grossklags, “Vulnerabilities of Data Protection in Vertical Federated Learning Training and Countermeasures”, IEEE Transactions on Information Forensics and Security, Vol 19, pp.3674-3689, 2024.
[27] H. Sun, Y. Zhang, Z. Xu, R. Zhang, M. Li, “MK-FLFHNN: A Privacy-Preserving Vertical Federated Learning Framework For Heterogeneous Neural Network Via Multi-Key Homomorphic Encryption”, 2023 26th International Conference on Computer Supported Cooperative Work in Design (CSCWD), 2023.
[28] 網路資料on line resources:FedAvg。取自https://biic.ee.nthu.edu.tw/blog-detail.php?id=2
[29] 網路資料on line resources:LUNA16數據集。取自https://biomedicalimaging.org/2016/?page_id=554
[30] N Nawreen, U Hany, T Islam “Lung Cancer Detection and Classification using CT Scan Image Processing”, 2021 International Conference on Automation, Control and Mechatronics for Industry 4.0 (ACMI), 2019.
[31] A. Sreekumar, K. R. Nair, S. Sudheer, H. G. Nayar; J. J. Nair “Malignant Lung Nodule Detection using Deep Learning”, 2020 International Conference on Communication and Signal Processing (ICCSP), 2020.
[32] A.Harsoyo, M. C. Rezi, P. H. Rusmin, “Design of face recognition system using local binary pattern and CLAHE on Smart Meeting Room System”, 2013 IEEE 3rd International Conference on System Engineering and Technology, 2013.
[33] L. Anifah, P. W. Rusimamto, Haryanto, S. I. Haryudo”Dentawyanjana Character Segmentation Using K-Means Clustering CLAHE Adaptive Thresholding Based”, 2022 Fifth International Conference on Vocational Education and Electrical Engineering (ICVEE),2022.
[34] 網路資料on line resources:侵蝕膨脹操作原理圖。取自https://hackmd.io/@cws0701/B1AxRjijq
[35] P. Sarker, M. M. H. Shuvo, Z. Hossain, S. Hasan “Segmentation and classification of lung tumor from 3D CT image using K-means clustering algorithm” , 2017 4th International Conference on Advances in Electrical Engineering (ICAEE), 2017.
[36] 網路資料on line resources:flower聯邦學習框架圖。取自https://flower.ai/docs/framework/tutorial-series-what-is-federated-learning.html
[37] P.Manju, B. R. Devassy, V.Rajan “A Novel Approach for Nuclei Segmentation Using U-Net”, 2023 International Conference on Networking and Communications (ICNWC), April 2023.

指導教授

吳中實(Jung-Shyr Wu)

審核日期

2024-7-15

推文