本研究旨在以ECG Lead I急性心肌梗塞辨識演算法建立兩階段急性心肌梗塞偵測的人工智慧模型，為此我們開發一套心肌梗塞心電圖標註系統及標準化的XML ECG資料解碼流程，以收集並提供醫師方便且簡潔的標註功能，讓醫生可以將醫院 心電儀輸出之心電圖在方便且簡潔的標註介面上將訊號的診斷結果以數位化的方式儲存，以為人工智慧模型的建模保留原始心電訊號與醫師判讀結果。本研究以上述系統共計收集2,367筆標註完成的心電圖，並挑選2,335筆12導程心電圖被標註為STEMI或Normal的ECG Lead I訊號做為AI模型的訓練及驗證樣本。透過訊號前處理篩選出1,360筆12導程心電訊號標記為STEMI的Lead I 10秒訊號和864筆12導程心電訊號標記為Normal的Lead I 10秒訊號。接著，將訊號分割成不同的心跳段落並使用波形描繪演算法來標記P、QRS和T波形的峰值位置、起始點和結束點。透過分析心跳波形上的能量強度，我們進行波段上的分割並從中萃取統計學特徵77種、QRS特徵15種、頻域特徵8種及小波域特徵44種，全部總共144種特徵。在第一階段中，我們的目標是訓練一個能夠具有泛化能力的心跳辨識模型，將資料按照心跳數目進行切割，並進行STEMI和Normal的二元分類。將12,731個STEMI心跳和8,162個Normal心跳進行10折分層交叉驗證，使用極限梯度提升模型(eXtreme Gradient Boosting, XGBoost)模型取得平均準確度、平均靈敏度、平均特異性及平均F1-score依序為93.5%、93.2%、93.2%、93%，模型辨識花費時間1.065秒。第二階段建立用於辨識心肌受損位置的模型，將6,017個前壁梗塞心跳、6,330個下壁梗塞心跳及384個側(後)壁梗塞心跳 進行10折分層交叉驗證，XGBoost模型取得平均準確度96.82%、平均靈敏度87.19%、平均特異性96.94%、平均F1-score 91%、模型辨識時間0.714秒。根據上述實驗結果，本研究完成以ECG Lead I建構之急性心肌梗塞AI模型之資料標註及演算法開發，未來將導入ECG心電手錶進行急性心肌梗塞偵測，讓患者能在日常生活中進行即時心臟健康的檢查，達到預防醫學中初級預防的效果。

The aim of this research i s to establish an artificial intelligence model for the detection of acute myocardial infarction using ECG Lead I-based algorithms. For this purpose, we developed a myocardial infarction electrocardiogram annotation system and a standardized XML-ECG data decoding process. This system allows physicians to conveniently and efficiently annotate ECG signals generated by hospital ECG machines on a user-friendly interface. The diagnosis results are then digitally stored, preserving the original ECG signals and the interpretations made by the physicians for the AI model's modeling.
In this study, a total of 2,367 annotated ECGs were collected using the annotation system. Out of these, 2,335 12-lead ECG signals were selected, with Lead I signals labeled as either STEMI or Normal, to serve as training and validation samples for the AI model. Signal preprocessing was performed to filter out 1,360 10-second Lead I signals labeled as STEMI, and 864 10-second Lead I signals labeled as Normal. Subsequently, the signals were segmented into different cardiac beats, and a waveform delineation algorithm was used to mark the peak positions, starting points, and ending points of P, QRS, and T waveforms. By analyzing the energy intensity of the heartbeat waveforms, we extracted a total of 144 statistical features, including 77 time-domain features, 15 QRS features, 8 frequency-domain features, and 44 wavelet-domain features.
In the first stage, the goal was to train a generalized heartbeat recognition model by segmenting the data based on the number of heartbeats and performing binary classification between STEMI and Normal. The XGBoost model was used, and 10-fold stratified cross-validation was applied to 12,731 STEMI heartbeats and 8,162 Normal heartbeats. The average accuracy, sensitivity, specificity, and F1-score achieved were 93.5%, 93.2%, 93.2%, and 93%, respectively, with a model recognition time of 1.065 seconds.
In the second stage, a model was established to identify the location of myocardial damage. 6,017 anterior myocardial infarction heartbeats, 6,330 inferior myocardial infarction, and 384 lateral (posterior) myocardial infarction heartbeats were used for 10-fold stratified cross-validation. The XGBoost model achieved an average accuracy of 96.82%, average sensitivity of 87.19%, average specificity of 96.94%, and average F1-score of 91%, with a model recognition time of 0.714 seconds.
Based on the experimental results, this study successfully completed the data annotation and algorithm development of the acute myocardial infarction AI model based on ECG Lead I. In the future, ECG smartwatches will be introduced for acute myocardial infarction detection, enabling patients to undergo real-time cardiac health checks in their daily lives, achieving the effect of primary prevention in preventive medicine.

[1] S. Yusuf, S. Hawken, and S. Ounpuu, “Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the Interheart Study): Case-control study,” ACC Current Journal Review, vol. 13, no. 12, pp. 15–16, 2004.
[2] “衛福部死因歷年統計, ” [Online]. Available: https://dep.mohw.gov.tw/DOS/lp-5069-113-xCat-y110.html (accessed Jun. 6, 2023).
[3] “Cardiovascular Disease,” World Health Organization, [Online]. Available: https://www.who.int/zh/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds) (accessed Jun. 6, 2023).
[4] Hui Yang, “Multiscale recurrence quantification analysis of Spatial Cardiac vectorcardiogram signals,” IEEE Transactions on Biomedical Engineering, vol. 58, no. 2, pp. 339–347, 2011.
[5] N. Herring and D. J. Paterson, “ECG diagnosis of acute ischaemia and infarction: past, present and future,” QJM Int. J. Med., vol. 99, no. 4, pp. 219–230, Apr. 2006.
[6] S. G. Ellis, M. Tendera, M. A. de Belder, A. J. van Boven, P. Widimsky, L. Janssens, H. R. Andersen, A. Betriu, S. Savonitto, J. Adamus, J. Z. Peruga, M. Kosmider, O. Katz, T. Neunteufl, J. Jorgova, M. Dorobantu, L. Grinfeld, P. Armstrong, B. R. Brodie, H. C. Herrmann, G. Montalescot, F.-J. Neumann, M. B. Effron, E. S. Barnathan, E. J. Topol, and FINESSE Investigators, “Facilitated PCI in patients with ST-elevation myocardial infarction,” N Engl J Med, vol. 358, no. 21, pp. 2205–2217, May 2008.
[7] I. Thylen, M. Ericsson, K. Hellstrom Angerud, R.-M. Isaksson, and S. Sederholm Lawesson, “First medical contact in patients with STEMI and its impact on time to diagnosis; an explorative cross-sectional study,” BMJ Open, vol. 5, no. 4, 2015.
[8] A. A. Alrumayh, A. M. Mubarak, A. A. Almazrua, M. Z. Alharthi, D. F. Alatef, T. B. Albacker, F. M. Samarkandy, Y. M. Alsofayan, and M. Alobaida, “Paramedic Ability in Interpreting Electrocardiogram with ST-segment Elevation Myocardial Infarction (STEMI) in Saudi Arabia,” J Multidiscip Healthc, vol. 15, pp. 1657–1665, 2022.
[9] G. Chandrashekar and F. Sahin, “A survey on feature selection methods,” Computers & Electrical Engineering, vol. 40, no. 1, pp. 16–28, 2014.
[10] P. Oktivasari, M. Hasyim, A. HS, F. H, and Suprijadi, “A simple real-time system for detection of normal and myocardial ischemia in the ST segment and T Wave Ecg Signal,” 2019 International Conference on Information and Communications Technology (ICOIACT), 2019.
[11] J. Liu, C. Zhang, Y. Zhu, T. Ristaniemi, T. Parviainen, and F. Cong, “Automated detection and localization system of myocardial infarction in single-beat ECG using Dual-Q TQWT and wavelet packet tensor decomposition,” Computer Methods and Programs in Biomedicine, vol. 184, p. 105120, Feb. 2020.
[12] F. A. Elhaj, N. Salim, A. R. Harris, T. T. Swee, and T. Ahmed, “Arrhythmia recognition and classification using combined linear and nonlinear features of ECG Signals,” Computer Methods and Programs in Biomedicine, vol. 127, pp. 52–63, 2016.
[13] S. Jain, M. K. Ahirwal, A. Kumar, V. Bajaj, and G. K. Singh, “QRS detection using Adaptive Filters: A comparative study,” ISA Transactions, vol. 66, pp. 362–375, 2017.
[14] M. B. Shahnawaz and H. Dawood, “An Effective Deep Learning Model for Automated Detection of Myocardial Infarction Based on Ultrashort-Term Heart Rate Variability Analysis,” Mathematical Problems in Engineering, vol. 2021, p. e6455053, Sep. 2021.
[15] U. R. Acharya, H. Fujita, M. Adam, O. S. Lih, V. K. Sudarshan, T. J. Hong, J. E. Koh, Y. Hagiwara, C. K. Chua, C. K. Poo, and T. R. San, “Automated characterization and classification of coronary artery disease and myocardial infarction by decomposition of ECG signals: A comparative study,” Information Sciences, vol. 377, pp. 17–29, Jan. 2017.
[16] A. Kaveh and W. Chung, “Automated classification of coronary atherosclerosis using single lead ECG,” in 2013 IEEE Conference on Wireless Sensor (ICWISE), pp. 108–113, Feb. 2013.
[17] J. Zhang, F. Lin, P. Xiong, H. Du, H. Zhang, M. Liu, Z. Hou, and X. Liu, “Automated Detection and Localization of Myocardial Infarction With Staked Sparse Autoencoder and TreeBagger,” IEEE Access, vol. 7, pp. 70634–70642, 2019.
[18] M. Kumar, R. B. Pachori, and U. R. Acharya, “Automated Diagnosis of Myocardial Infarction ECG Signals Using Sample Entropy in Flexible Analytic Wavelet Transform Framework,” Entropy, vol. 19, no. 9, Art. no. 9, Sep. 2017.
[19] P. Kora and S. R. Kalva, “Improved Bat algorithm for the detection of myocardial infarction,” SpringerPlus, vol. 4, no. 1, p. 666, Nov. 2015.
[20] K. Thygesen, J. S. Alpert, A. S. Jaffe, B. R. Chaitman, J. J. Bax, D. A. Morrow, H. D. White, and ESC Scientific Document Group, “Fourth universal definition of myocardial infarction (2018),” European Heart Journal, vol. 40, no. 3, pp. 237–269, Jan. 2019.
[21] P. J. Zimetbaum and M. E. Josephson, “Use of the electrocardiogram in acute myocardial infarction,” New England Journal of Medicine, vol. 348, no. 10, pp. 933–940, 2003.
[22] K. Thygesen, J. S. Alpert, and H. D. White, “Universal definition of myocardial infarction,” Journal of the American College of Cardiology, vol. 50, no. 22, pp. 2173–2195, 2007.
[23] A.L. Goldberger, Z. D. Goldberger, and A. Shvilkin, Clinical Electrocardiography: A Simplified Approach E-Book, Elsevier Health Sciences, Amsterdam, Netherlands, 2017.
[24] N. A. Bhaskar, “Performance analysis of Support Vector Machine and neural networks in detection of myocardial infarction,” Procedia Computer Science, vol. 46, pp. 20–30, 2015.
[25] K. Jafarian, V. Vahdat, S. Salehi, and M. Mobin, “Automating detection and localization of myocardial infarction using shallow and end-to-end deep neural networks,” Applied Soft Computing, vol. 93, p. 106383, 2020.
[26] M. Arif, I. A. Malagore, and F. A. Afsar, “Detection and localization of myocardial infarction using K-nearest neighbor classifier,” Journal of Medical Systems, vol. 36, no. 1, pp. 279–289, 2010.
[27] U. R. Acharya, H. Fujita, M. Adam, O. S. Lih, V. K. Sudarshan, T. J. Hong, J. E. Koh, Y. Hagiwara, C. K. Chua, C. K. Poo, and T. R. San, “Automated characterization and classification of coronary artery disease and myocardial infarction by decomposition of ECG signals: A comparative study,” Information Sciences, vol. 377, pp. 17–29, Jan. 2017.
[28] G. B. Berikol, O. Yildiz, and İ. T. Özcan, “Diagnosis of acute coronary syndrome with a support vector machine,” Journal of Medical Systems, vol. 40, no. 4, 2016.
[29] A. K. Dohare, V. Kumar, and R. Kumar, “Detection of myocardial infarction in 12 lead ECG using support Vector Machine,” Applied Soft Computing, vol. 64, pp. 138–147, 2018.
[30] H. M. Rai and K. Chatterjee, “Hybrid CNN-LSTM deep learning model and ensemble technique for automatic detection of myocardial infarction using BIG ECG Data,” Applied Intelligence, vol. 52, no. 5, pp. 5366–5384, 2021.
[31] U. B. Baloglu, M. Talo, O. Yildirim, R. S. Tan, and U. R. Acharya, “Classification of myocardial infarction with multi-lead ECG signals and deep CNN,” Pattern Recognition Letters, vol. 122, pp. 23–30, 2019.
[32] M. Hammad, M. H. Alkinani, B. B. Gupta, and A. A. Abd El-Latif, “Myocardial infarction detection based on deep neural network on imbalanced data,” Multimedia Systems, vol. 28, no. 4, pp. 1373–1385, 2021.
[33] X. Lu, H. Duan, and H. Zheng, “XML-ECG: An XML-based ECG presentation for data exchanging,” 2007 1st International Conference on Bioinformatics and Biomedical Engineering, 2007.
[34] R. R. Bond, D. D. Finlay, C. D. Nugent, and G. Moore, “A review of ECG storage formats,” International Journal of Medical Informatics, vol. 80, no. 10, pp. 681–697, 2011.
[35] E. Kalai Zaghouani, A. Benzina, and R. Attia, “ECG based authentication for e-healthcare systems: Towards a secured ECG features transmission,” 2017 13th International Wireless Communications and Mobile Computing Conference (IWCMC), 2017.
[36] U. Satija, B. Ramkumar, and M. Sabarimalai Manikandan, “Real-time signal quality-aware ECG telemetry system for IOT-based health care monitoring,” IEEE Internet of Things Journal, vol. 4, no. 3, pp. 815–823, 2017.
[37] 曾文慶, 蔣家正, 鄭涵菁, 莊文南, 陳朝慶, “以 XML 為基礎的心電圖管理系統, ”Journal of Information Technology and Applications (資訊科技與應用期刊), vol. 1, no. 2, pp. 115-121, 2006.
[38] Rahul Kher, “Signal Processing Techniques for Removing Noise from ECG Signals,” Journal of Biomedical Engineering and Research, Dec. 2019.
[39] D. Thanapatay, C. Suwansaroj, and C. Thanawattano, “ECG beat classification method for ECG printout with principle components analysis and support vector machines,” 2010 International Conference on Electronics and Information Engineering, 2010.
[40] V. X. Afonso, W. J. Tompkins, T. Q. Nguyen, and Shen Luo, “ECG beat detection using filter banks,” IEEE Transactions on Biomedical Engineering, vol. 46, no. 2, pp. 192–202, 1999.
[41] M. Aqil, A. Jbari, and A. Bourouhou, “ECG signal  denoising by discrete wavelet transform,” International Journal of Online Engineering (iJOE), vol. 13, no. 09, p. 51, 2017.
[42] R. J. Martis, U. R. Acharya, and L. C. Min, “ECG beat classification using PCA, Lda, ICA and discrete wavelet transform,” Biomedical Signal Processing and Control, vol. 8, no. 5, pp. 437–448, 2013.
[43] M. K. Sarkaleh, “Classification of ECG arrhythmias using discrete wavelet transform and Neural Networks,” International Journal of Computer Science, Engineering and Applications, vol. 2, no. 1, pp. 1–13, 2012.
[44] “離散小波變換,”[Online]. Available: https://zh.wikipedia.org/zh-tw/離散小波變換 (accessed Jun. 15, 2023).
[45] 陳冠宇, “改良小波去雜訊門檻方法於心電圖訊號,” 碩士論文, 輔仁大學資訊工程學系, 2009.
[46] C. Cai and P. de Harrington, “Different discrete wavelet transforms applied to denoising analytical data,” Journal of Chemical Information and Computer Sciences, vol. 38, no. 6, pp. 1161–1170, 1998.
[47] Li Su and Guoliang Zhao, “De-noising of ECG signal using translation- invariant wavelet de-noising method with improved thresholding,” 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference, 2005.
[48] S. Poornachandra, “Wavelet-based denoising using subband dependent threshold for ECG signals,” Digital Signal Processing, vol. 18, no. 1, pp. 49–55, 2008.
[49] L. Kramer, C. Menon, and M. Elgendi, “ECGAssess: A python-based toolbox to assess ECG lead signal quality,” Frontiers in Digital Health, vol. 4, 2022.
[50] J. Pan and W. J. Tompkins, “A real-time QRS detection algorithm,” IEEE Transactions on Biomedical Engineering, vol. BME-32, no. 3, pp. 230–236, 1985.
[51] W. Zong, G. B. Moody, and D. Jiang, “A robust open-source algorithm to detect onset and duration of QRS complexes,” Computers in Cardiology, 2003.
[52] M. I. Sezan, “A peak detection algorithm and its application to histogram-based Image Data Reduction,” Computer Vision, Graphics, and Image Processing, vol. 47, no. 3, p. 396, 1989.
[53] J. P. Martinez, R. Almeida, S. Olmos, A. P. Rocha, and P. Laguna, “A wavelet-based ECG delineator: evaluation on standard databases,” IEEE Transactions on Biomedical Engineering, vol. 51, no. 4, pp. 570–581, Apr. 2004.
[54] S. Banerjee, R. Gupta, and M. Mitra, “Delineation of ECG characteristic features using multiresolution wavelet analysis method,” Measurement, vol. 45, no. 3, pp. 474–487, Apr. 2012.
[55] H.-Y. Lin, S.-Y. Liang, Y.-L. Ho, Y.-H. Lin, and H.-P. Ma, “Discrete-wavelet-transform-based noise removal and feature extraction for ECG signals,” IRBM, vol. 35, no. 6, pp. 351–361, Dec. 2014.
[56] S. Banerjee, “A First Derivative Based R-Peak Detection and DWT Based Beat Delineation Approach of Single Lead Electrocardiogram Signal,” in 2019 IEEE Region 10 Symposium (TENSYMP), pp. 565–570, Jun. 2019.
[57] I. Beraza and I. Romero, “Comparative study of algorithms for ECG segmentation,” Biomedical Signal Processing and Control, vol. 34, pp. 166–173, Apr. 2017.
[58] N. Safdarian, N. J. Dabanloo, and G. Attarodi, “A New Pattern Recognition Method for Detection and Localization of Myocardial Infarction Using T-Wave Integral and Total Integral as Extracted Features from One Cycle of ECG Signal,” Journal of Biomedical Science and Engineering, vol. 2014, Aug. 2014.
[59] M. Arif, I. A. Malagore, and F. A. Afsar, “Automatic Detection and Localization of Myocardial Infarction Using Back Propagation Neural Networks,” in 2010 4th International Conference on Bioinformatics and Biomedical Engineering, pp. 1–4, Jun. 2010.
[60] A. V. Bharadwaj, S. M. Upadhyaya, S. L., and R. Srinivasan, “Early Diagnosis and Automated Analysis of Myocardial Infarction (STEMI) by Detection of ST Segment Elevation Using Wavelet Transform and Feature Extraction,” in 2018 International Conference on Design Innovations for 3Cs Compute Communicate Control (ICDI3C), pp. 24–28, Apr. 2018.
[61] L. Sun, Y. Lu, K. Yang, and S. Li, “ECG Analysis Using Multiple Instance Learning for Myocardial Infarction Detection,” IEEE Transactions on Biomedical Engineering, vol. 59, no. 12, pp. 3348–3356, Feb. 2012.
[62] W. Zhang, R. Li, S. Shen, J. Yao, Y. Peng, G. Chen, B. Zhou, and Z. Wang, “Interpretable Detection and Location of Myocardial Infarction Based on Ventricular Fusion Rule Features,” J Healthc Eng, vol. 2021, p. 4123471, 2021.
[63] H. Liu, M. Zhou, and Q. Liu, “An embedded feature selection method for Imbalanced Data Classification,” IEEE/CAA Journal of Automatica Sinica, vol. 6, no. 3, pp. 703–715, 2019.
[64] A. Jovic, K. Brkic, and N. Bogunovic, “A review of feature selection methods with applications,” 2015 38th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), 2015.
[65] Y. Han, J. Wu, B. Zhai, Y. Pan, G. Huang, L. Wu, and W. Zeng, “Coupling a bat algorithm with XGBoost to estimate reference evapotranspiration in the arid and semiarid regions of China,” Advances in Meteorology, vol. 2019, pp. 1–16, 2019.
[66] H. Vazirani, A. Shukla, R. Kala, and R. Tiwari, “Diagnosis of breast cancer by Modular Neural Network,” 2010 3rd International Conference on Computer Science and Information Technology, 2010.
[67] R. Bousseljot, D. Kreiseler, and A. Schnabel, "Nutzung der EKG-Signaldatenbank CARDIODAT der PTB über das Internet," Biomedizinische Technik, vol. 40, suppl. 1, pp. 317, 1995.
[68] G. B. Moody and R. G. Mark, "The impact of the MIT-BIH Arrhythmia Database," IEEE Engineering in Medicine and Biology, vol. 20, no. 3, pp. 45-50, May-June 2001.
[69] “驗證策略,” [Online]. Available: https://toutiao.io/posts/apx38n7/preview (accessed Jun. 21, 2023).
[70] B. Fatimah, P. Singh, A. Singhal, D. Pramanick, P. S., and R. B. Pachori, “Efficient detection of myocardial infarction from single lead ECG signal,” Biomedical Signal Processing and Control, vol. 68, p. 102678, Jul. 2021.
[71] S. Khatun and B. I. Morshed, “Detection of myocardial infarction and arrhythmia from single-lead ECG data using bagging trees classifier,” 2017 IEEE International Conference on Electro Information Technology (EIT), 2017.
[72] C. M. Gibson, S. Mehta, M. R. S. Ceschim, A. Frauenfelder, D. Vieira, R. Botelho, F. Fernandez, C. Villagran, S. Niklitschek, C. I. Matheus, G. Pinto, I. Vallenilla, C. Lopez, M. I. Acosta, A. Munguia, C. Fitzgerald, J. Mazzini, L. Pisana, and S. Quintero, “Evolution of single-lead ECG for STEMI detection using a deep learning approach,” International Journal of Cardiology, vol. 346, pp. 47–52, Jan. 2022.
[73] L.-M. Tseng, C.-Y. Chuang, S.-K. Chua, and V. S. Tseng, “Identification of Coronary Culprit Lesion in ST Elevation Myocardial Infarction by Using Deep Learning,” IEEE Journal of Translational Engineering in Health and Medicine, vol. 11, pp. 70–79, 2023.
[74] L. Wu, G. Huang, X. Yu, M. Ye, L. Liu, Y. Ling, X. Liu, D. Liu, B. Zhou, Y. Liu, J. Zheng, S. Liang, R. Pu, X. He, Y. Chen, L. Han, and X. Qian, “Deep Learning Networks Accurately Detect ST-Segment Elevation Myocardial Infarction and Culprit Vessel,” Frontiers in Cardiovascular Medicine, vol. 9, 2022.
[75] Z. Bouzid, Z. Faramand, R. E. Gregg, S. Helman, C. Martin-Gill, S. Saba, C. Callaway, E. Sejdić, and S. Al-Zaiti, “Novel ECG features and machine learning to optimize culprit lesion detection in patients with suspected acute coronary syndrome,” Journal of Electrocardiology, vol. 69, pp. 31–37, Nov. 2021.