臺灣博碩士論文加值系統

摘要 i
ABSTRACT ii
誌謝 iii
目錄 iv
表目錄 vii
圖目錄 viii
第一章 緒論 1
1.1 前言 1
1.2 研究動機 2
1.3 論文架構 4
第二章　　背景知識 5
2.1 非侵入式血糖量測之國內外研究 5
2.1.1 使用機器學習對智慧型手機 PPG 訊號進行血糖值迴歸 5
2.1.2 基於雙波段PPG和生物電阻抗量測神經網路的非侵入性血糖估算系統 6
2.1.3 使用機器學習實現非侵入式血糖量測 7
2.1.4 穿戴式裝置應用於非侵入量測血液資訊及模型建立之研究 8
2.2 PPG光譜感測技術 9
2.3 深度神經網路模型框架 9
2.3.1 VGG16&VGG19 9
2.3.2 GoogLeNet 11
2.3.3 ResNet 13
2.3.4 EfficientNet 13
2.3.5 Vision Transformer 15
第三章　　研究方法 17
3.1 研究架構 17
3.2 非侵入式穿戴式血糖量測系統 17
3.2.1 系統架構 18
3.2.2 微型光譜感測晶片 NSP32 18
3.2.3 行動裝置應用程式 19
3.3 IRB試驗 21
3.3.1 IRB及人體試驗介紹 21
3.3.2 資料採集標準作業流程 22
3.3.3 實驗裝置安裝 23
3.3.4 實驗數據分析及標記 25
3.4 訊號處理 26
3.4.1 實驗數據資料庫 26
3.4.2 光譜訊號處理 27
3.4.3 熱力圖建立 27
3.5 分析方法 29
3.5.1 混淆矩陣 29
3.5.2 ROC曲線及AUC曲線下面積 31
第四章　　實驗與結果 33
4.1 非侵入式穿戴式血糖量測系統 33
4.1.1 非侵入式穿戴式血糖量測裝置 33
4.2 IRB試驗數據分析 35
4.2.1 血糖項目人次分布表 36
4.3 深度神經網路模型訓練與測試 40
4.3.1 VGG16&VGG19模型建立及測試結果 40
4.3.1.1 VGG16模型同一醫院資料測試結果 42
4.3.1.2 VGG16模型不同醫院資料測試結果 46
4.3.1.3 VGG19模型同一醫院資料測試結果 50
4.3.1.4 VGG19模型不同醫院資料測試結果 54
4.3.2 GoogLeNet模型建立及測試結果 58
4.3.2.1 GoogLeNet模型同一醫院資料測試結果 58
4.3.2.2 GoogLeNet模型不同醫院資料測試結果 62
4.3.3 ResNet模型建立及測試結果 66
4.3.3.1 ResNet模型同一醫院資料測試結果 67
4.3.3.2 ResNet模型不同醫院資料測試結果 71
4.3.4 EfficientNet模型建立及測試結果 76
4.3.4.1 EfficientNet模型同一醫院資料測試結果 77
4.3.4.2 EfficientNet模型不同醫院資料測試結果 81
4.3.5 VisionTransformer模型建立及測試結果 85
4.3.5.1 VisionTransformer模型同一醫院資料測試結果 85
4.3.5.2 VisionTransformer模型不同醫院資料測試結果 89
4.3.6 各模型表現比較 93
第五章　　結論與未來展望 96
5.1 結論 96
5.2 未來展望 97
參考文獻 98

[1].“認識糖尿病 : 原因、症狀、診斷、預防”, [Online] Available:
https://helloyishi.com.tw/diabetes/what-is-diabetes/. [Accessed: Jun. 18, 2024].
[2].“衛生福利部國民健康署 2017-2020 年「國民營養健康狀況變遷調查」” [Online]
Available: https://www.hpa.gov.tw/Pages/List.aspx?nodeid=3998. [Accessed: Jun, 2024].
[3].“衛生福利部 控糖5撇步 健康有保固” [Online] Available: https://www.mohw.gov.tw/cp-5020-63343-1.html. [Accessed: Jun, 2024].
[4].“衛生福利部國民健康署 控糖五招，降低糖尿病併發症保平安！” [Online]
Available: https://www.hpa.gov.tw/Pages/Detail.aspx?nodeid=4705&pid=17623.
[Accessed: Jun, 2024].
[5].“國家發展委員會扶養比趨勢” [Online] Available:
https://www.ndc.gov.tw/Content_List.aspx?n=695E69E28C6AC7F3.
[Accessed: Jun, 2024].
[6].T. T. Islam, et al., "Blood Glucose Level Regression for Smartphone PPG Signals Using Machine Learning," Applied Sciences, vol. 11, no. 2, p. 618, 2021, https://doi.org/10.3390/app11020618.
[7].Chien-Tai Yen, et al., "Non-Invasive Blood Glucose Estimation System Based on a Neural Network with Dual-Wavelength Photoplethysmography and Bioelectrical Impedance Measuring," Sensors, vol. 22, no. 12, p. 4452, 2022, https://doi.org/10.3390/s22124452.
[8].Shantanu Sen Gupta, et al., "Towards Non-Invasive Blood Glucose Measurement Using Machine Learning: An All-Purpose PPG System Design," Biomedical Signal Processing and Control, vol. 68, p. 102706, 2021, https://doi.org/10.1016/j.bspc.2021.102706.
[9].張翔峻。「穿戴式裝置應用於非侵入量測血液資訊及模型建立之研究」。碩士論文，國立臺北科技大學電機工程, 2023. https://hdl.handle.net/11296/67pgn3.
[10].K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition," arXiv preprint arXiv:1409.1556, 2015.
[11].C. Szegedy, et al., "Going Deeper with Convolutions," 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 2015, pp. 1-9, doi: 10.1109/CVPR.2015.7298594.
[12].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), Las Vegas, NV, USA, 2016, pp. 770-778, doi: 10.1109/CVPR.2016.90.
[13].M. Tan and Q. V. Le, "EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks," arXiv preprint arXiv:1905.11946, 2019.
[14].Alexey Dosovitskiy, et al., "An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale," arXiv preprint arXiv:2010.11929, 2020.
[15].Endang Susana, et al., "Non-Invasive Classification of Blood Glucose Level for Early Detection Diabetes Based on Photoplethysmography Signal," Information, vol. 13, no. 2, p. 59, 2022, https://doi.org/10.3390/info13020059.
[16].“nanoLambda官方網站”[Online] Available: https://nanolambda.myshopify.com/. [Accessed: Jun, 24, 2024].
[17].“NSP32m datasheet” [Online] Available:
http://165.227.7.198/Resources/Release/Data%20Sheet_NSP32m_v1.0.0.pdf. [Accessed: Jan, 20, 2023].
[18].“研究倫理委員會申請說明” [Online] Available: https://www.yongxi-stat.com/irb/. [Accessed: Jun. 24, 2024].
[19].“[Data Science] 什麼是混淆矩陣 (Confusion Matrix) -模型評估指標” [Online] Available: https://tako-analytics.com/2024-03-21-data-science-what-is-confusion-matrix-model-evaluation-metric/. [Accessed: Jun, 24, 2024].
[20].“深入介紹及比較ROC曲線及PR曲線” [Online] Available: https://medium.com/nlp-tsupei/roc-pr-%E6%9B%B2%E7%B7%9A-f3faa2231b8c. [Accessed: Jun, 24, 2024].
[21].“[論文導讀] Vision Transformer (ViT) 附程式碼實作” [Online] Available: https://medium.com/@andy6804tw/%E8%AB%96%E6%96%87%E5%B0%8E%E8%AE%80-vision-transformer-vit-%E9%99%84%E7%A8%8B%E5%BC%8F%E7%A2%BC%E5%AF%A6%E4%BD%9C-379306ea2fb. [Accessed: Jun, 24, 2024].
[22].“論文閱讀: GoogLeNet/Inception(Going deeper with convolutions)” [Online] Available: https://hackmd.io/@RinHizakura/HJxhXSbE8. [Accessed: Jun, 24, 2024].
[23].S.-H. Chen, Y.-C. Chuang, and C.-C. Chang, "Development of a Portable All-Wavelength PPG Sensing Device for Robust Adaptive-Depth Measurement: A Spectrometer Approach with a Hydrostatic Measurement Example," Sensors, vol. 20, no. 22, p. 6556, 2020, https://doi.org/10.3390/s20226556.
[24].Anju Prabha, et al., "Intelligent Estimation of Blood Glucose Level Using Wristband PPG Signal and Physiological Parameters," Biomedical Signal Processing and Control, vol. 78, p. 103876, 2022, https://doi.org/10.1016/j.bspc.2022.103876.
[25].Ö. Pektaş and M. Köseoğlu, "Common AI-Based Methods Used in Blood Glucose Estimation with PPG Signals," In Smart Applications with Advanced Machine Learning and Human-Centred Problem Design, ICAIAME 2021, Engineering Cyber-Physical Systems and Critical Infrastructures, vol. 1, Springer, Cham, 2023, https://doi.org/10.1007/978-3-031-09753-9_44.
[26].Sajid Ahmed, et al., "Blood Glucose Level Regression for Smartphone PPG Signals Using Machine Learning," Applied Sciences, vol. 11, no. 2, p. 618, 2021, https://doi.org/10.3390/app11020618.