Responsive image
博碩士論文 etd-0621120-203312 詳細資訊
Title page for etd-0621120-203312
論文名稱
Title
深度學習應用於十二導程心電圖病徵分類之研究
Applying Deep Learning to Classification of 12-lead Electrocardiography Symptoms
系所名稱
Department
畢業學年期
Year, semester
語文別
Language
學位類別
Degree
頁數
Number of pages
70
研究生
Author
指導教授
Advisor
召集委員
Convenor
口試委員
Advisory Committee
口試日期
Date of Exam
2020-07-23
繳交日期
Date of Submission
2020-07-21
關鍵字
Keywords
兩階段式學習、多尺度網路、深度學習、異常檢測、十二導程心電圖
Multi-scale network, 12-lead Electrocardiography, anomaly detection, deep learning, two-phase learning
統計
Statistics
本論文已被瀏覽 5728 次,被下載 0
The thesis/dissertation has been browsed 5728 times, has been downloaded 0 times.
中文摘要
十二導程心電圖是醫院在心臟疾病中常用的輔助診斷指標,心電圖的檢測由於其價格低、無侵入的特性被廣泛用於心臟疾病的篩檢、診查及體檢中,且每天的檢測量巨大。目前多導程的心電圖設備已經廣泛用於臨床當中,部分設備已經具有自動分析診斷功能,但自動分析對於多種心電圖異常事件的判別還不夠精確,需要醫生做進一步確認。
近年來人工智慧在心電圖的分類及預測領域有了許多應用,深度學習的技術發展有望助力心電圖波形、心電圖異常事件的分類及預測,從而達到提升預測精度的目標。本研究藉由高雄醫學大學附設醫院的研究數據庫所提供之十二導程心電圖,並由心臟內科的專業醫師進行標記,在標記的過程中為了避免誤判,利用了異常檢測的方式排除及修正。
在此研究的數據庫中,由於心電圖類別標記的筆數非常不平衡,而且各個病徵類別包含了許多不同的特徵,因此本研究建立了兩階段式的多尺度深度學習模型,運用兩階段式的學習,在第一階段可以將數量較多的類別區分出來,避免在訓練的過程中產生過度凝合,以至於結果偏向於數量多的那一類,因此先將資料分為是否有病徵兩類進行訓練;第二階段再將第一階段分類有病徵的資料,輸入多尺度深度學習模型,利用多尺度深度學習模型的特性,進行不同尺度的縮放,可以提取不同尺度的特徵進行病徵分類,使用該模型建立十二導程心電圖判讀分類系統,最後結果顯示,在精確率及召回率中分別得到95.49%及98.31%,在F1-Measure可得到96.88%。
Abstract
The 12-lead Electrocardiography (ECG) is a commonly used auxiliary diagnostic indicator in hospital for heart disease. Electrocardiography detection is widely used in screening, diagnosis, and physical examination of heart diseases due to its low price and non-invasive characteristics. And the daily detection volume is huge. At present, multi-lead ECG equipment has been widely used in clinical. Some devices already have automatic analysis and diagnosis functions. However, the automatic analysis is not accurate enough to discriminate against many abnormal ECG events. The doctor needs further confirmation.
In recent years, artificial intelligence has many applications in the field of ECG classification and prediction. The development of deep learning technology is expected to help the classification and prediction of ECG waveform and ECG abnormal events. To achieve the goal of improving prediction accuracy. This study was based on the 12-lead ECG provided by the research database of the Kaohsiung Medical University Hospital (KMUH). And marked by professional doctor in cardiology. To avoid misjudgment during the marking process, the method of anomaly detection is used to eliminate and correct.
In the database of this study, since the number of ECG category markers was very unbalanced. And each symptom category contains many different characteristics. Thus, this study established a two-stage multi-scale deep learning model. Using two-stage learning, in the first stage, a large number of categories can be distinguished to avoid over-fitting during training. So that the result is biased towards the larger number. And so, the information is divided into two types for training. In the second stage, the data with symptoms from the first stage, input to multi-scale deep learning model. Using the characteristics of multi-scale deep learning models, scaling at different scales can extract features of different scales to classify symptoms. Use this model to establish a 12-lead ECG interpretation classification system. The result shows that precision rate and recall rates are 95.49% and 98.31% respectively, and F1-Measure can get 96.88%.
目次 Table of Contents
目錄
論文審定書 i
誌謝 ii
摘要 iii
Abstract iv
圖目錄 vii
表目錄 viii
第一章 導論 1
1.1. 研究背景 1
1.2. 研究動機與目的 3
1.3. 心臟傳導系統 4
1.4. 十二導程心電圖概述 5
1.5. 十二導程心電圖資料數據 6
1.5.1. 高醫的心電圖資料及性別年齡數據 6
1.5.2. 中國生理訊號挑戰賽公開資料庫 8
1.5.3. 將高醫的研究數據庫及中國生理訊號挑戰賽公開資料庫混合 9
1.6. 論文架構 9
1.7. 論文貢獻 10
第二章 文獻探討 11
第三章 研究方法 16
3.1. 相關前置處理 17
3.1.1. 心電圖資料數位化 17
3.1.2. 醫生進行心電圖標記 21
3.1.3. 去除標記中的異常資料 23
3.2. 深度學習演算法介紹 26
3.2.1. 卷積神經網路 26
3.2.2. 多尺度卷積神經網路 29
3.2.3. 循環神經網路 30
3.2.4. 卷積循環神經網路 32
3.2.5. 兩階段式的多尺度深度學習模型 33
3.3. 心電訊號特徵分析 36
3.4. 本章小結 44
第四章 實驗結果 45
4.1. 實驗測量指標 45
4.2. 高醫的研究數據庫所提供之十二導程心電圖 46
4.3. 中國生理訊號挑戰賽公開資料庫 50
4.4. 將高醫的研究數據庫及挑戰賽公開資料庫混合訓練 51
4.5. 本章小結 52
第五章 結論與未來展望 53
5.1. 結論 53
5.2. 未來研究方向 53
參考文獻 54
參考文獻 References
[1] Einthoven's triangle https://en.wikipedia.org/wiki/Einthoven%27s_triangle
[2] Electrocardiograph CardiMax FCP-8100 https://fukuda.com/products/fcp-8100/
[3] 高雄醫院研究資料庫 Kaohsiung Medical University Hospital Research Database (KMUHRD)
[4] The China Physiological Signal Challenge 2018 http://2018.icbeb.org/Challenge.html
[5] Liu, F., Liu, C., Zhao, L., Zhang, X., Wu, X., Xu, X., & Li, J. (2018). An open access database for evaluating the algorithms of electrocardiogram rhythm and morphology abnormality detection. Journal of Medical Imaging and Health Informatics, 8(7), 1368-1373.
[6] OpenCV https://opencv.org/
[7] Chandola, V., Banerjee, A., & Kumar, V. (2009). Anomaly detection: A survey. ACM computing surveys (CSUR), 41(3), 1-58.
[8] Chalapathy, R., & Chawla, S. (2019). Deep learning for anomaly detection: A survey. arXiv preprint arXiv:1901.03407.
[9] Wang, Z. Y., Wu, C. Y., Lin, Y. T., & Lee, S. J. (2019). Weighted z-Distance-Based Clustering and Its Application to Time-Series Data. Applied Sciences, 9(24), 5469.
[10] Hannun, A. Y., Rajpurkar, P., Haghpanahi, M., Tison, G. H., Bourn, C., Turakhia, M. P., & Ng, A. Y. (2019). Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network. Nature medicine, 25(1), 65.
[11] Li, Y., Zhang, Y., Zhao, L., Zhang, Y., Liu, C., Zhang, L., ... & Li, J. (2018). Combining convolutional neural network and distance distribution matrix for identification of congestive heart failure. IEEE Access, 6, 39734-39744.
[12] Sun, W., Zeng, N., & He, Y. (2019). Morphological arrhythmia automated diagnosis method using gray-level co-occurrence matrix enhanced convolutional neural network. IEEE Access, 7, 67123-67129.
[13] Li, F., Xu, Y., Chen, Z., & Liu, Z. (2019). Automated heartbeat classification using 3-D inputs based on convolutional neural network with multi-fields of view. IEEE Access, 7, 76295-76304.
[14] Huang, J., Chen, B., Yao, B., & He, W. (2019). ECG arrhythmia classification using STFT-based spectrogram and convolutional neural network. IEEE Access, 7, 92871-92880.
[15] Kiranyaz, S., Ince, T., & Gabbouj, M. (2015). Real-time patient-specific ECG classification by 1-D convolutional neural networks. IEEE Transactions on Biomedical Engineering, 63(3), 664-675.
[16] Liu, W., Zhang, M., Zhang, Y., Liao, Y., Huang, Q., Chang, S., ... & He, J. (2017). Real-time multilead convolutional neural network for myocardial infarction detection. IEEE journal of biomedical and health informatics, 22(5), 1434-1444.
[17] Zhang, Q., Zhou, D., & Zeng, X. (2017). HeartID: A multiresolution convolutional neural network for ECG-based biometric human identification in smart health applications. Ieee Access, 5, 11805-11816.
[18] Saadatnejad, S., Oveisi, M., & Hashemi, M. (2019). LSTM-based ECG classification for continuous monitoring on personal wearable devices. IEEE journal of biomedical and health informatics, 24(2), 515-523.
[19] Li, R., Zhang, X., Dai, H., Zhou, B., & Wang, Z. (2019). Interpretability analysis of heartbeat classification based on heartbeat activity’s global sequence features and BiLSTM-attention neural network. IEEE Access, 7, 109870-109883.
[20] He, R., Liu, Y., Wang, K., Zhao, N., Yuan, Y., Li, Q., & Zhang, H. (2019). Automatic cardiac arrhythmia classification using combination of deep residual network and bidirectional LSTM. IEEE Access, 7, 102119-102135.
[21] Shashikumar, S. P., Shah, A. J., Clifford, G. D., & Nemati, S. (2018, July). Detection of paroxysmal atrial fibrillation using attention-based bidirectional recurrent neural networks. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 715-723).
[22] Ballinger, B., Hsieh, J., Singh, A., Sohoni, N., Wang, J., Tison, G. H., ... & Pletcher, M. J. (2018, April). DeepHeart: semi-supervised sequence learning for cardiovascular risk prediction. In Thirty-Second AAAI Conference on Artificial Intelligence.
[23] Hong, S., Xiao, C., Ma, T., Li, H., & Sun, J. (2019). Mina: Multilevel knowledge-guided attention for modeling electrocardiography signals. arXiv preprint arXiv:1905.11333.
[24] Doersch, C. (2016). Tutorial on variational autoencoders. arXiv preprint arXiv:1606.05908.
[25] Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., ... & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
[26] Lin, T. Y., Goyal, P., Girshick, R., He, K., & Dollár, P. (2017). Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision (pp. 2980-2988).
[27] Wang, Y., Yao, Q., Kwok, J. T., & Ni, L. M. (2020). Generalizing from a few examples: A survey on few-shot learning. ACM Computing Surveys (CSUR), 53(3), 1-34.
[28] Golany, T., & Radinsky, K. (2019, July). PGANs: Personalized generative adversarial networks for ECG synthesis to improve patient-specific deep ECG classification. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 33, pp. 557-564).
[29] Zhou, B., Liu, S., Hooi, B., Cheng, X., & Ye, J. (2019, August). BeatGAN: Anomalous Rhythm Detection using Adversarially Generated Time Series. In IJCAI (pp. 4433-4439).
[30] Jiang, J., Zhang, H., Pi, D., & Dai, C. (2019). A novel multi-module neural network system for imbalanced heartbeats classification. Expert Systems with Applications: X, 1, 100003.
[31] Rajesh, K. N., & Dhuli, R. (2018). Classification of imbalanced ECG beats using re-sampling techniques and AdaBoost ensemble classifier. Biomedical Signal Processing and Control, 41, 242-254.
[32] Liu, X. Y., Wu, J., & Zhou, Z. H. (2008). Exploratory undersampling for class-imbalance learning. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 39(2), 539-550.
[33] Amin, A., Anwar, S., Adnan, A., Nawaz, M., Howard, N., Qadir, J., ... & Hussain, A. (2016). Comparing oversampling techniques to handle the class imbalance problem: A customer churn prediction case study. IEEE Access, 4, 7940-7957.
[34] Riaz, S., Arshad, A., & Jiao, L. (2018). Rough noise-filtered easy ensemble for software fault prediction. Ieee Access, 6, 46886-46899.
[35] Liu, X. Y., & Zhou, Z. H. (2013). Ensemble methods for class imbalance learning. Imbalanced Learning: Foundations, Algorithms and Applications, 61-82.
[36] Bao, L., Juan, C., Li, J., & Zhang, Y. (2016). Boosted Near-miss Under-sampling on SVM ensembles for concept detection in large-scale imbalanced datasets. Neurocomputing, 172, 198-206.Mousavi, S., Afghah, F., & Acharya, U. R. (2019).
[37] Mousavi, S., Afghah, F., & Acharya, U. R. (2019). SleepEEGNet: Automated sleep stage scoring with sequence to sequence deep learning approach. PloS one, 14(5), e0216456.
[38] Mousavi, S., & Afghah, F. (2019, May). Inter-and intra-patient ecg heartbeat classification for arrhythmia detection: a sequence to sequence deep learning approach. In ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 1308-1312). IEEE.
[39] Mousavi, S., Afghah, F., Razi, A., & Acharya, U. R. (2019, May). ECGNET: Learning where to attend for detection of atrial fibrillation with deep visual attention. In 2019 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI) (pp. 1-4). IEEE.
[40] Cai, J., Sun, W., Guan, J., & You, I. (2020). Multi-ECGNet For ECG Arrythmia Multi-Label Classification. IEEE Access.
[41] Mei, Q., & Liu, T. (2020). ECG diagnosis based on one-dimensional convolutional neural network. MS&E, 768(7), 072110.
[42] Seo, W., Kim, N., Kim, S., Lee, C., & Park, S. M. (2019). Deep ECG-respiration network (DeepER net) for recognizing mental stress. Sensors, 19(13), 3021.
[43] Chen, T. M., Huang, C. H., Shih, E. S., Hu, Y. F., & Hwang, M. J. (2020). Detection and Classification of Cardiac Arrhythmias by a Challenge-Best Deep Learning Neural Network Model. Iscience, 23(3), 100886.
[44] Cai, W., & Hu, D. (2020). QRS complex detection using novel deep learning neural networks. IEEE Access.
[45] He, R., Liu, Y., Wang, K., Zhao, N., Yuan, Y., Li, Q., & Zhang, H. (2019). Automatic cardiac arrhythmia classification using combination of deep residual network and bidirectional LSTM. IEEE Access, 7, 102119-102135.
[46] Liu, Z., Meng, X. A., Cui, J., Huang, Z., & Wu, J. (2018, October). Automatic identification of abnormalities in 12-lead ecgs using expert features and convolutional neural networks. In 2018 International Conference on Sensor Networks and Signal Processing (SNSP) (pp. 163-167). IEEE.
[47] Wang, R., Yao, Q., Fan, X., & Li, Y. (2019, October). Multi-class arrhythmia detection based on neural network with multi-stage features fusion. In 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) (pp. 4082-4087). IEEE.
[48] Qiu, L., Cai, W., Yu, J., Zhong, J., Wang, Y., Li, W., ... & Wang, L. (2020). A two-stage ECG signal denoising method based on deep convolutional network. bioRxiv.
[49] Sokolova, M., & Lapalme, G. (2009). A systematic analysis of performance measures for classification tasks. Information processing & management, 45(4), 427-437.
[50] Tensorflow https://www.tensorflow.org/
[51] Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
[52] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
[53] Sajid, S., Hussain, S., & Sarwar, A. (2019). Brain tumor detection and segmentation in MR images using deep learning. Arabian Journal for Science and Engineering, 44(11), 9249-9261.
[54] Yang, J., Li, W., Wang, R., Xue, L., & Hu, M. (2019). Enhanced two-phase residual network for single image super-resolution. Journal of Visual Communication and Image Representation, 61, 188-197.
[55] Lai, C. T., Hong, Y. T., Chen, H. Y., Lu, C. J., & Lin, S. D. (2019, November). Multiple text style transfer by using word-level conditional generative adversarial network with two-phase training. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) (pp. 3570-3575).
[56] https://en.wikipedia.org/wiki/Convolutional_neural_network
[57] https://en.wikipedia.org/wiki/K-means_clustering
[58] Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).
[59] Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2016). Inception-v4, inception-resnet and the impact of residual connections on learning. arXiv preprint arXiv:1602.07261.
電子全文 Fulltext
本電子全文僅授權使用者為學術研究之目的,進行個人非營利性質之檢索、閱讀、列印。請遵守中華民國著作權法之相關規定,切勿任意重製、散佈、改作、轉貼、播送,以免觸法。
論文使用權限 Thesis access permission:自定論文開放時間 user define
開放時間 Available:
校內 Campus:開放下載的時間 available 2025-07-21
校外 Off-campus:開放下載的時間 available 2025-07-21

您的 IP(校外) 位址是 44.204.24.82
現在時間是 2024-03-29
論文校外開放下載的時間是 2025-07-21

Your IP address is 44.204.24.82
The current date is 2024-03-29
This thesis will be available to you on 2025-07-21.

紙本論文 Printed copies
紙本論文的公開資訊在102學年度以後相對較為完整。如果需要查詢101學年度以前的紙本論文公開資訊,請聯繫圖資處紙本論文服務櫃台。如有不便之處敬請見諒。
開放時間 available 2025-07-21

QR Code