Improved Two-layer BiLSTM Electrocardiosignal Segmentation Method

doi:10.12068/j.issn.1005-3026.2023.12.005

Abstract

Abstract: The segmentation method of ECG signals can effectively reflect athlete’s heart function and physical performance. Manual segmentation of ECG signals often consumes a lot of time and energy. In order to achieve automated ECG signal segmentation， this paper proposes improved two-layer BiLSTM electrocardiosignal segmentation algorithm， which can analyze time series forward and backward to detect and locate important waveforms such as the P-wave， QRS complex， and T-wave. The experiment used a publicly available QT database to simulate pre-competition ECG data of athletes. In comparative experiments with traditional LSTM， BiLSTM and two-layer BiLSTM， all evaluation indicators of this method have been improved. The average accuracy rate is 95.68%， the average recall rate is 91.62%， the average precision is 91.05%， the average specificity is 96.64%， and the average F1 score is 91.41%. The results showed that the proposed method has good performance in ECG signal segmentation.

Key words: long short-term memory; exercise electrocardiogram; signal segmentation; recurrent neural network; deep learning

CLC Number:

YANG Xuan， HE Zhan-qi. Improved Two-layer BiLSTM Electrocardiosignal Segmentation Method[J]. Journal of Northeastern University(Natural Science), 2023, 44(12): 1705-1711.

References

[1]中国心血管健康与疾病报告编写组.中国心血管健康与疾病报告2020概要［J］.中国循环杂志，2021，36(6):521-545.(The Writing Committee of the Report on Cardiovascular Health and Diseases in China.Report on cardiovascular health and diseases burden in China:an updated summary of 2020［J］.Chinese Circulation Journal，2021，36(6):521-545.)
[2]Abrishami H，Campbell M，Han C，et al.P-QRS-T localization in ECG using deep learning［C］//2018 IEEE EMBS International Conference on Biomedical & Health Informatics.Las Vegas:IEEE-Verlag，2018:210-213.
[3]Lay-Ekuakille A，Vergallo P，Griffo G，et al.Entropy index in quantitative EEG measurement for diagnosis accuracy［J］.IEEE Transactions on Instrumentation and Measurement，2014，63(6):1440-1450.
[4]Wen X，Huang Y M，Shen T H，et al.Prevalence of abnormal and borderline electrocardiogram changes in 13 079 Chinese amateur marathon runners［J］.BMC Sports Science Medicine and Rehabilitation，2021，13:41.
[5]Kiranyaz S，Ince T，Gabbouj M.Real-time patient-specific ECG classification by 1-D convolutional neural networks［J］.IEEE Transactions on Biomedical Engineering，2016，63(3):664-675.
[6]Li P F，Wang Y，He J C，et al.High-performance personalized heartbeat classification model for long-term ECG signal［J］.IEEE Transactions on Biomedical Engineering，2017，64(1):78-86.
[7]Tran L，Li Y F，Nocera L，et al.MultiFusionNet:atrial fibrillation detection with deep neural networks［C］// AMIA Summits on Translational Science Proceedings.Bethesda:AMIA-Verlag，2020:654-663.
[8]Kamozawa H，Muroga S，Tanaka M.A detection method of atrial fibrillation from 24-hour holter-ECG using CNN［J］.IEEJ Transactions on Electrical and Electronic Engineering，2023，18(4):577-582.
[9]Che C，Zhang P L，Zhu M，et al.Constrained transformer network for ECG signal processing and arrhythmia classification［J］.BMC Medical Informatics and Decision Making，2021，21(1):184.
[10]Malali A，Hiriyannaiah S，Siddesh G M，et al.Supervised ECG wave segmentation using convolutional LSTM［J］.ICT Express，2020，6(3):166-169.
[11]Londhe A N，Atulkar M.Semantic segmentation of ECG waves using hybrid channel-mix convolutional and bidirectional LSTM［J］.Biomedical Signal Processing and Control，2021，63:102162.
[12]Atiea M A，Refaat H E.A review on machine，transfer and deep learning approaches for ECG classification［J］.Frontiers in Scientific Research and Technology，2023，5(1):18-28.
[13]陈琦玲，吴寸草，陈琪.运动员心电图解析的国际建议［J］.临床心电学杂志，2018，27(5):321-334.(Chen Qi-ling，Wu Cun-cao，Chen Qi.International suggestions on EGG analysis of athletes［J］.Journal of Clinical Electrocardiology，2018，27(5):321-334.)
[14]Cheng J Y，Zou Q X，Zhao Y X.ECG signal classification based on deep CNN and BiLSTM［J］.BMC Medical Informatics and Decision Making，2021，21(1):365.
[15]Maglaveras N，Stamkopoulos T，Pappas C，et al.ECG processing techniques based on neural networks and bidirectional associative memories［J］.Journal of Medical Engineering & Technology，1998，22(3):106-111.
[16]Liu X L，Deng Z D，Yang Y H.Recent progress in semantic image segmentation［J］.Artificial Intelligence Review，2019，52:1089-1106.
[17]Yang L，Xie K，Wen C，et al.Speech emotion analysis of netizens based on bidirectional LSTM and PGCDBN［J］.IEEE Access，2021，9:59860-59872.
[18]Dohare A K，Kumar V，Kumar R.An efficient new method for the detection of QRS in electrocardiogram［J］.Computers and Electrical Engineering，2014，40(5):1717-1730.
[19]Hochreiter S，Schmidhuber J.Long short-term memory［J］.Neural Computation，1997，9(8):1735-1780.
[20]Graves A，Jaitly N，Mohamed A.Hybrid speech recognition with deep bidirectional LSTM［C］//Workshop on Automatic Speech Recognition and Understanding.Olomouc:IEEE-Verlag，2013:273-278.
[21]Srivastava N，Hinton G，Krizhevsky A，et al.Dropout:a simple way to prevent neural networks from overfitting［J］.Journal of Machine Learning Research，2014，15:1929-1958.
[22]Laguna P，Mark R G，Goldberg A，et al.A database for evaluation of algorithms for measurement of QT and other waveform intervals in the ECG［C］//Computers in Cardiology 1997.Lund:IEEE-Verlag，1997，24:673-676.