臺灣博碩士論文加值系統

每年有許多人死於病痛之下，因生病造成死亡的比例中，心血管疾病就佔了多數部分，絕大部分醫生使用聽診器來辨識病患心音是否正異常，但是實際上心音的診斷很大程度上依賴醫生的臨床經驗和檢查技巧，訓練有素的醫生也不見得能辨識出一樣的答案，所以開發出心音辨識的演算法，有效辨識正常及異常心音，可以在第一時間提供醫生們臨床參考數據，讓醫生有一定的把握確認患者是否為異常心音，使患者能更早獲得正確的治療。
此研究計畫分為心音週期中的類別辨識演算法和心音週期投票演算法兩部分，(1)使用ECGdeli開源工具箱、LR-HSMM、Bi-LSTM模型求出心音週期中的各類別，過程中使用特徵包絡和K-fold等手法將訊號有效地做處理，最後辨識出第一心音、心音收縮期、第二心音、心音舒張期。(2)將辨識過的心音，做週期性的分割，使用Bi-LSTM和SVDD作為模型基礎，對於每個心音週期，使用時域的統計量和頻域的梅爾頻率倒譜係數做為特徵，之後使用投票的方法，將心音以特定投票比例，分類辨識正異常。經過實驗，投票率在5~30%時，辨識準確度92.85%、靈敏度89.44%、特異度93.71%和F1 Score 83.74%。

Every year, many people die from illnesses, and cardiovascular diseases account for most of the deaths caused by diseases. Most doctors use stethoscopes to identify whether a patient's heart sound is abnormal or not, but in fact, the diagnosis of heart sounds relies heavily on the clinical experience and examination skills of doctors, and even well-trained doctors may not be able to identify the same answer. Therefore, the development of a heart sound classificatiton algorithm to effectively identify normal and abnormal heart sounds can provide doctors with clinical reference data in the first instance, allowing them to confirm with certainty whether a patient has an abnormal heart sound, so that the patient can receive correct treatment earlier.
This research project is divided into two parts: the classification algorithm in heart sound cycle and the heart sound cycle voting algorithm. (1) The ECGdeli open source toolbox, LR-HSMM, and Bi-LSTM models were used to find out the various stages of heart sound cycles, and the signals were processed effectively using feature envelopes and K-folds. Finally, the first heart sound, systolic heart sound, second heart sound, and diastolic heart sound were identified. (2) The classified heart sounds were segmented periodically using Bi-LSTM and SVDD as the basis of the model, and for each heart sound cycle, the time domain statistics and the frequency domain inverse coefficient of Meier frequency were used as the features. Then, the heart sounds were categorized as normal and abnormal using a voting method with a specific voting ratio. Then, the heart sounds were categorized as normal and abnormal using a voting method with a specific voting ratio. After the experiments, the accuracy of classification was 92.85%, the sensitivity was 89.44%, the specificity was 93.71%, and the F1 Score was 83.74% when the voting rate ranged from 5 to 30%.

致謝 i
摘要 ii
ABSTRACT iii
目錄 iv
圖目錄 vii
表目錄 xi
第一章 緒論 1
1.1研究動機與目的 1
1.2文獻探討 3
1.3研究方法 6
1.4論文架構 8
第二章 心電圖標記心音位置 10
2.1心電圖與心音圖的關聯性 10
2.2使用心電圖標註點預測心音類別 18
2.3 特徵選取 18
2.4 HSMM模型 20
2.5 邏輯回歸 22
第三章 心音分割演算法 25
3.1分幀 25
3.2 K-fold 26
3.3 特徵處理 26
3.4 LSTM模型 27
3.5 Bi-LSTM模型 29
3.6 整體架構 30
第四章 心音週期投票演算法 31
4.1特徵處理 32
4.2 模型使用介紹 34
4.3 整體架構 37
4.4 投票方式 39
第五章 實驗結果 40
5.1 樣本來源 40
5.2 評分方式 41
5.3 心音分割結果 42
5.4 心音週期排列 48
5.5 Bi-LSTM模型心音週期切割結果 51
5.6 SVDD模型心音週期切割結果 57
5.7 實驗討論 60
第六章 結論 64
參考文獻 65
附錄 實驗結果 71

[1] WHO心血管疾病文獻，取自網路:
https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)
[2] S. K. Gardezi, S. G. Myerson, J. Chambers, S. Coffey, J. d’Arcy, et al., “Cardiac auscultation poorly predicts the presence of valvular heart disease in asymptomatic primary care patients”, Heart, vol. 104, pp. 1832-1835. May, 2018.
[3] 心臟雜音相關資訊:
https://doctor.get.com.tw/m/Journal/detail.aspx?no=406299
[4] M. Morshed, S. A. Fattah and M. Saquib, “Automated Heart Valve Disorder Detection Based on PDF Modeling of Formant Variation Pattern in PCG Signal”, IEEE Access, vol. 10, pp. 27330-27342, March, 2022.
[5] S. Sun, H. Wang, C. Cheng, Z. Chang and D. Huang, “PCA-based heart sound feature generation for a ventricular septal defect discrimination”, 2017 14th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP), 2017, pp. 128-133.
[6] C. D. Papadaniil and L. J. Hadjileontiadis, “Efficient Heart Sound Segmentation and Extraction Using Ensemble Empirical Mode Decomposition and Kurtosis Features”, in IEEE Journal of Biomedical and Health Informatics, vol. 18, no. 4, pp. 1138-1152, July, 2014.
[7] M. Deng, T. Meng, J. Cao, S. Wang, J. Zhang, & H. Fan, (2020). “Heart sound classification based on improved MFCC features and convolutional recurrent neural networks”, Neural Networks, vol. 130, pp. 22-32. October, 2020.
[8] H. Liang, S. Lukkarinen and I. Hartimo, “Heart sound segmentation algorithm based on heart sound envelogram,” Computers in Cardiology 1997, 1997, pp. 105-108.
[9] Q. Liu, X. Wu, & X. Ma, “An automatic segmentation method for heart sounds”, Biomedical engineering online, vol. 17, no. 106, July, 2018.
[10] A. Bouril, D. Aleinikava, M. S. Guillem and G. M. Mirsky, “Automated classification of normal and abnormal heart sounds using support vector machines”, 2016 Computing in Cardiology Conference (CinC), 2016, pp. 549-552.
[11] F. A. Khan, A. Abid, M. S. Khan. “Automatic heart sound classification from segmented/unsegmented phonocardiogram signals using time and frequency features”, Physiological measurement, vol. 41, no. 5, June, 2020.
[12] D. Kumar, P. Carvalho, M. Antunes, R. P. Paiva and J. Henriques, “Heart murmur classification with feature selection”, 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, 2010, pp. 4566-4569.
[13] V. Arora, R. Leekha, R. Singh, & I. Chana, “Heart sound classification using machine learning and phonocardiogram”, Modern Physics Letters B, vol. 33, no. 26. August, 2019.
[14] T. Fernando, H. Ghaemmaghami, S. Denman, S. Sridharan, N. Hussain and C. Fookes, “Heart Sound Segmentation Using Bidirectional LSTMs With Attention”, in IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 6, pp. 1601-1609. October, 2019.
[15] M. Bahreini, R. Barati, & A. Kamaly, (2022). “Heart Sound Classification Based on Fractal Dimension and MFCC Features Using Hidden Markov Model”, January, 2022.
[16] S. I. Khan and V. Ahmed, “Classification of pulmonary crackles and pleural friction rubs using MFCC statistical parameters”, 2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2016, pp. 2437-2440.

[17] M. Rahmandani, H. A. Nugroho and N. A. Setiawan, “Cardiac Sound Classification Using Mel-Frequency Cepstral Coefficients (MFCC) and Artificial Neural Network (ANN)”, 2018 3rd International Conference on Information Technology, Information System and Electrical Engineering (ICITISEE), 2018, pp. 22-26.
[18] D. Kurniadi, Y. Kung, Z. Chen, Y. Li, Hendrick and G. Jong, “Implemented the expert system of heart disease by using SVM”, 2018 IEEE International Conference on Applied System Invention (ICASI), 2018, pp. 605-608.
[19] S. A. Singh, & S. Majumder, “Classification of unsegmented heart sound recording using KNN classifier”, Journal of Mechanics in Medicine and Biology, vol. 19, no. 04, June, 2019.
[20] M. Zubair, “A Peak Detection Algorithm for Localization and Classification of Heart Sounds in PCG Signals using K-means Clustering”, August, 2021.
[21] Z. Tariq, S. K. Shah, & Y. Lee, “Feature-based Fusion using CNN for Lung and Heart Sound Classification”, Sensors, vol.22, no. 4, December, 2021.
[22] C. Lubis, & F. Gondawijaya, “Heart sound diagnose system with BFCC, MFCC, and backpropagation neural network”, In IOP conference series: materials science and engineering, vol. 508, no. 1, pp. 012119. April, 2019.
[23] C. Liu, D. Springer, Q. Li, B. Moody, R. A. Juan, F. J. Chorro, F. Castells, J. M. Roig, I. Silva, A. E. Johnson, and Z. Syed, “An open access data base for the evaluation of heart sound algorithms”, Physiological measurement, vol. 37, no. 12, pp. 2181-2213. December, 2016.
[24] A. Goldberger, L. Amaral, L. Glass, J. Hausdorff, P. C. Ivanov, R. Mark, ... & H. E. Stanley, “PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals”, Circulation, vol. 101, no. 23, pp. e215-e220. June, 2000.
[25] D. B. Springer, L. Tarassenko and G. D. Clifford, “Logistic Regression-HSMM-Based Heart Sound Segmentation”, in IEEE Transactions on Biomedical Engineering, vol. 63, no. 4, pp. 822-832. April, 2016.
[26] P. Laguna, R. Jané, and P. Caminal, “Automatic detection of wave boundaries in multilead ECG signals: validation with the CSE database”, Computers and Biomedical Research, vol. 27, 1994, pp. 45-60. February, 1994.
[27] A. Demski, & M. L. Soria, “ecg-kit: a Matlab toolbox for cardiovascular signal processing”, Journal of open research software, vol. 4, no. 1, April, 2016.
[28] A. N. Vest, G. D. Poian, Q. Li, C. Liu, S. Nemati, A. J. Shah, & G. D. Clifford, “An open source benchmarked toolbox for cardiovascular waveform and interval analysis”, Physiological measurement, vol. 39, no. 10, October, 2018.
[29] N. Pilia, C. Nagel, G. Lenis, S. Becker, O. Dössel, A. Loewe, “ECGdeli - An open source ECG delineation toolbox for MATLAB”, SoftwareX, Vol. 13, January, 2021.
[30] 心電圖週期標註點示意圖，取自網路:
https://reurl.cc/k1x60x
[31] G. Lenis, N. Pilia, T. Oesterlein, A. Luik, C. Schmitt, and O. Dössel, “P wave detection and delineation in the ECG based on the phase free stationary wavelet transform and using intracardiac atrial electrograms as reference”, Biomedical Engineering / Biomedizinische Technik, vol. 61, no. 1, pp. 37-56. February, 2016.
[32] 基於邏輯回歸的隱藏半馬爾可夫心音分割工具箱:
https://physionet.org/content/hss/1.0/
[33] D. Gill, N. Gavrieli and N. Intrator, “Detection and identification of heart sounds using homomorphic envelogram and self-organizing probabilistic model”, Computers in Cardiology, 2005, 2005, pp. 957-960.
[34] A. Thalmayer, S. Zeising, G. Fischer, J. Kirchner, “A Robust and Real-Time Capable Envelope-Based Algorithm for Heart Sound Classification: Validation under Different Physiological Conditions”, Sensors, vol. 20, no. 4, February, 2020.
[35] H. Tang, M. Wang, Y. Hu, B. Guo, & T. Li, “Automated Signal Quality Assessment for Heart Sound Signal by Novel Features and Evaluation in Open Public Datasets”, BioMed Research International, 2021, February, 2021.
[36] Liang Huiying, L. Sakari and H. Iiro, “A heart sound segmentation algorithm using wavelet decomposition and reconstruction”, Proceedings of the 19th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. 'Magnificent Milestones and Emerging Opportunities in Medical Engineering' (Cat. No.97CH36136), 1997, pp. 1630-1633.
[37] S. E. Schmidt, et al., “Segmentation of heart sound recordings by a duration-dependent hidden Markov model.” Physiological Measurement, vol. 31, no. 4, pp. 513-29, April, 2010.
[38] L. Rabiner, “A tutorial on hidden Markov models and selected applications in speech recognition,” Proceedings of the IEEE, vol. 77, no. 2, pp. 257-286, February, 1989.
[39] T. Masuko, K. Tokuda, T. Kobayashi and S. Imai, “Voice characteristics conversion for HMM-based speech synthesis system,” 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing, 1997, pp. 1611-1614.
[40] J. A. Venkidasalapathy, & C. Kravaris, “Hidden markov model based approach for diagnosing cause of alarm signals”, AIChE Journal, vol.67, no. 10, May, 2021.
[41] S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory”, in Neural Computation, vol. 9, no. 8, pp. 1735-1780, November, 1997.
[42] A. J. Gaona, P. D. Arini, “Deep Recurrent Learning for Heart Sounds Segmentation based on Instantaneous Frequency Features”, Elektron: ciencia y tecnología en la electrónica de hoy, vol. 4, no. 2, pp. 52-57, October, 2020.
[43] 香濃熵，取自網路:
https://baike.baidu.hk/item/%E9%A6%99%E8%BE%B2%E7%86%B5/1649961
[44] 譜熵，取自網路:
https://baike.baidu.com/item/%E8%B0%B1%E7%86%B5/22689242
[45] 梅爾頻率倒譜係數介紹，取自網路:
https://zh.wikipedia.org/wiki/梅爾頻率倒譜係數
[46] D. M. Tax, R. P. Duin, “Support Vector Data Description”, Machine Learning, vol. 54, pp.45–66, January, 2004.
[47] Y. Zhao, S. Wang, F. Xiao, “Pattern recognition-based chillers fault detection method using Support Vector Data Description (SVDD)”, Applied Energy, Vol. 112, pp. 1041-1048, December, 2013.
[48] G. D. Clifford et al., “Classification of normal/abnormal heart sound recordings: The PhysioNet/Computing in Cardiology Challenge 2016”, 2016 Computing in Cardiology Conference (CinC), 2016, pp. 609-612.
[49] S. C. Oliveira, E. F. Gomes, and A. M. Jorge, “Heart sounds classification using motif based segmentation”, In Proceedings of the 18th International Database Engineering & Applications Symposium, 2014, pp. 370-371.
[50] M. Zabihi, A. B. Rad, S. Kiranyaz, M. Gabbouj and A. K. Katsaggelos, “Heart sound anomaly and quality detection using ensemble of neural networks without segmentation”, 2016 Computing in Cardiology Conference (CinC), 2016, pp. 613-616.
[51] E. Kay and A. Agarwal, “DropConnected neural network trained with diverse features for classifying heart sounds”, 2016 Computing in Cardiology Conference (CinC), 2016, pp. 617-620.