近年來，因為科技的日新月異，鄉村的人口會漸漸移至城市讓城市的人口密集提高，在人口密集的情況下最需的就是管理人民的健康，所以在已成熟的物聯網(Internet of Things)的架構上，各種產品連上指定的雲端做不同的監控以及多樣化的服務，而這樣的架構用於在醫療領域非常不普及，反而這幾年來在業界才開始被重視，醫療領域套用物聯網架構就會轉變成智慧醫療(Smart Health)，智慧醫療不只是終端的醫療設備上傳數據到指定的醫療雲端，還需要深入探討醫療雲端的管理架構、智能的數據分析、遠距長照、急救服務……等，本研究針對的議題是心肌梗塞，首先會先建置ECG穿戴裝置讓病患可以隨時上傳ECG的數據到醫療雲端，在醫療雲端的架構分兩種1.醫療網頁平台建置提供病患醫師的使用 2.數據分析提供醫師診斷時可參考依據，當偵測到心肌梗塞異常警報的發生時讓急救小組能在最快速的時間查看病患資料及聯繫，再決定是不是需要派車的動作，心肌梗塞的分析使用捲積神經網路(Convolutional Neural Networks)進行訓練，不只能分析心肌梗塞或是健康的分類，也能提供5種不同部位心肌梗塞的分類器，提供醫師診斷建議。
未來的醫療平台不再只是讓民眾單純的掛號，醫生也不用主觀的診斷治療，從終端設備到雲端架構再到數據分析都將會是新的架構，在人工智慧發達的年代，數據分析會是智慧醫療的主力，急救的服務也會越來越好，讓人民享有更好的醫療服務的體制。

In recent years, with the rapid advancement of science and technology, the rural population will gradually move to urban areas to make the urban population densely populated. In the case of dense population, the most needed is to manage people's health, Therefore, in the mature structure of Internet of Things, various products connect to the designated cloud for different monitoring and diversified services, and such an architecture is not widely used in the Smart Health, On the contrary, in the past few years, the industry has begun to be taken seriously.
The application of the Internet of Things framework in the medical field will be transformed into Smart Healthcare. Smart medical equipment not only uploads data to the designated medical cloud on the terminal's medical device, but also probes into the medical cloud Management structure, intelligent data analysis, remote healthcare, emergency services... and so on.
This study addresses the issue of myocardial infarction, first to set up ECG wearable device allows patients to upload ECG data to the medical cloud at any time, in the medical cloud architecture have two ways: 1. Medical website platform to provide the patient and doctor use 2. Data analysis can provide reference for the diagnosis of physicians, emergency service decision exception alarm occurs when the first aid team can view the patient data and contacts in the fastest time, and then decide whether the need to send an ambulance.

Myocardial infarction analysis, which uses Convolutional Neural Networks for training, not only analyzes myocardial infarction or health classifications, but also provides a classifier for five different positions of myocardial infarction, providing physician diagnostic advice.
The future work of the medical platform is no longer just let the people simply register, the doctor also do not subjective diagnosis and treatment, from the terminal equipment to the cloud architecture and then to the data analysis will be the new architecture, in the age of developed artificial intelligence, data analysis will be the main force of smart medical services, emergency services will be better and better, so that people have better medical services system.

摘要 iii
ABSTRACT iv
誌 謝 vi
目錄 vii
表目錄 ix
圖目錄 x
第一章、緒論 1
1.1研究動機 1
1.2研究目的 3
1.3論文架構 5
第二章、文獻探討 6
2.1淺談心肌梗塞 6
2.2心電圖介紹 6
2.3智慧醫療體系架構 7
2.4相關文獻 8
2.5特性比較 9
第三章、系統架構 10
3.1 硬體介紹 10
3.2軟體開發環境介紹 10
3.3系統架構 10
3.4 系統流程 11
3.5心肌梗塞訓練模型及流程 14
3.5.1訓練流程 14
3.5.2心肌梗塞捲積神經網路架構 15
3.6急救服務的決策 15
第四章、系統實現與評估 16
4.1系統建構 16
4.2模擬情境 16
4.3心肌梗塞的捲積神經網路分析結果 18
4.3.1分析結果呈現方式 18
4.3.2心肌梗塞模型分析結果 19
第五章、結論與未來規劃 22
參考資料 23

[1] NHLBI. (2014). What Are the Signs and Symptoms of Coronary Heart Disease? [Online]. Available: www.nhlbi.nih.gov. [Accessed 28 October 2017].
[2] Gartner, Inc.(2016). Gartner's 2016 Hype Cycle for Emerging Technologies Identifies Three Key Trends That Organizations Must Track to Gain Competitive Advantage. [Online]. Available: https://www.gartner.com/newsroom/id/3412017 [Accessed 28 October 2017].
[3] TRACTICA.(2017). Wearable Device Shipments to Reach 430 Million Units Annually by 2022. [Online]. Available: https://goo.gl/e42NPf . [Accessed 15 October 2017].
[4] 台灣衛生福利統計處.(2016).104年死因統計結果分析. [Online]. Available: http://dep.mohw.gov.tw/DOS/lp-1777-113.html. [Accessed 05 August 2016].
[5] 台灣衛生福利統計處.(2017). 105年死因結果摘要表. [Online]. Available: http://dep.mohw.gov.tw/DOS/lp-3352-113.html.[ Accessed 04 August 2017]
[6] NHLBI. (2014). Heart Attack.[Online].Available: https://www.nhlbi.nih.gov/health-topics/heart-attack.[ Accessed 28 October 2017].
[7] America Heart Association.(2016).What Are the Signs and Symptoms of Coronary Heart Disease?. [Online]. Available :http://www.heart.org/HEARTORG/Conditions/More/MyHeartandStrokeNews/Heart-Attack-or-Sudden-Cardiac-Arrest-How-Are-They-Different_UCM_440804_Article.jsp#.Wm_TbaiWa70.[ Accessed 04 February 2017].
[8] 亞東紀念醫院.常見心血管疾.[Online]. Available : http://www.femhcv.com.tw/des_04.php[Accessed 04 February 2017].
[9] Nobelprize.org.(2002).The electrocardiogram – looking at the heart of electricity.[Online]. Available : https://www.nobelprize.org/educational/medicine/ecg/ecg-readmore.html.
[10] YANOWITZ, Frank G. ECG learning center. University of Utah School of Medicine,[Online]. Available: http://ecg. utah. edu/.[Accessed 27 October 2017].
[11] 陳文彬、潘祥林，診斷學（第8版），人民衛生出版社。(2013).pp.120、486、487.
[12] Akmandor, Ayten Ozge, and Niraj K. Jha. "Smart Health Care: An Edge-Side Computing Perspective." IEEE Consumer Electronics Magazine, vol. 7, no. 1, pp. 29-37, Jan. 2018.
[13] KOVALCHUK, Sergey V., et al. Distributed data-driven platform for urgent decision making in cardiological ambulance control. Future Generation Computer Systems, vol.79, no. 1, pp.144-154, February 2018.
[14] FARAHANI, Bahar, et al. Towards fog-driven IoT eHealth: Promises and challenges of IoT in medicine and healthcare. Future Generation Computer Systems, vol. 78, no. 2, pp.659-676, January 2018.
[15] HOSSAIN, M. Shamim; MUHAMMAD, Ghulam. Cloud-assisted industrial internet of things (iiot)–enabled framework for health monitoring. Computer Networks, vol. 101, no. 4, pp.192-202, June 2016.
[16] LIU, Wenhan, et al. Real-time Multilead Convolutional Neural Network for Myocardial Infarction Detection. IEEE Journal of Biomedical and Health Informatics, vol. PP, no. 99, pp.1-1 , November 2017.
[17] ZIHLMANN, Martin; PEREKRESTENKO, Dmytro; TSCHANNEN, Michael. Convolutional Recurrent Neural Networks for Electrocardiogram Classification. arXiv preprint arXiv:1710.06122, November 2017.
[18] Yin, Shihui, et al. "Designing ECG-based physical unclonable function for security of wearable devices." Engineering in Medicine and Biology Society (EMBC), 2017 39th Annual International Conference of the IEEE. IEEE, September 2017.
[19] Physionet. The PTB Diagnostic ECG Database.(2004). Available: https://physionet.org/physiobank/database/ptbdb/.
[20] Stehman, Stephen V. "Selecting and interpreting measures of thematic classification accuracy." Remote sensing of Environment, vol. 62, no. 1, pp. 77-89, October 1997
[21] Deeks, Jonathan J., and Douglas G. Altman. "Diagnostic tests 4: likelihood ratios." Bmj , vol. 27, July 2004.