T-Wave Morphological Classification Based on CNN and Modified Frequency Slice Wavelet Transform
Xie Jiajing1, Wei Shoushui1*, Jiang Xinge2#, Wang Chunyuan1, Cui Huaijie1, Liu Chengyu3#
1(School of Control Science and Engineering, Shandong University, Jinan 250061, China) 2(School of Information Science and Engineering, Shandong Jiaotong University, Jinan 250357, China) 3(State Key Laboratory of Bioelectronics, School of Instrument Science and Engineering, Southeast University, Nanjing 210096,Jiangsu, China)
Abstract：Real-time monitoring of ECG is one important means of cardiovascular disease prevention. T-wave is an important characteristic of diseases such as myocardial ischemia and sudden cardiac death. The automatic identification of T-wave is a challenging taskin ECG remote monitoring. Due to the influence of high noise background of real time monitoring ECG, the conventional T-wave feature extraction and classification algorithm encounters a bottleneck. In this paper, a T-wave morphological recognition algorithm combining slice frequency wavelet transform and convolutional neural network was proposed. The algorithm included: locating automatically the R-wave peak's position and the T-wave ends' position to identify a segment containing the T-wave; the frequency slice wavelet transform was performed, and the generated time-frequency image was input into the convolutional neural network to complete the classification of the T-wave. The frequency slice wavelet transform transformed the signal to the time-frequency domain, which accurately presented the time-frequency energy distribution characteristics of the ECG signal. The hidden layers of the convolutional neural network completed the three features' extraction of the time-frequency image by convolving, activating and pooling the time-frequency image three times. These features have translation and scaling invariance. In this paper, 12 830 fragments in European ST-T database was used. The convolutional neural network model was trained and tested by the 3-fold cross validation method. The classification accuracy of experiment based on heart beats reached 97.34%, and the F1 measure reached 96.97%. The classification accuracy of experiment based on samples was 84.80%, and the F1 measure was 83.29%. The classification accuracy of the model tested in QT database was 87.83%, F1 measure was 85.38%, and the generalization performance was good. Compared with other T-wave classification algorithms (such as decision tree, support vector machine, etc.), the classification accuracy based on heart beat experiments was improved by 1～5%. The results demonstrated that the algorithm designed for the classification of six types of T-wave improved the accuracy and performed well in terms of robustness and generalization performance. In addition, the algorithm model was also applicable to the analysis of other physiological signals and has certain guiding significance in the field of medical image analysis.
谢佳静, 魏守水, 江兴娥, 王春元, 崔怀杰, 刘澄玉. 基于CNN和频率切片小波变换的T波形态分类[J]. 中国生物医学工程学报, 2021, 40(1): 1-11.
Xie Jiajing, Wei Shoushui, Jiang Xinge, Wang Chunyuan, Cui Huaijie, Liu Chengyu. T-Wave Morphological Classification Based on CNN and Modified Frequency Slice Wavelet Transform. Chinese Journal of Biomedical Engineering, 2021, 40(1): 1-11.
[1] 操端, 王海滨, 胡玉良. 基于智能手机的远程心电显示系统的设计与研究 [J]. 西华大学学报(自然科学版), 2009, 28(4): 16-19. [2] 刘澄玉, 杨美程, 邸佳楠, 等. 穿戴式心电:发展历程、核心技术与未来挑战 [J]. 中国生物医学工程学报, 2019, 38(6): 641-652. [3] 夏云龙, 杨延宗. T波形成的心电生理机制及其争议 [J]. 心血管病学进展, 2010, 31(4): 13-16.) [4] 钱敏. 心电图T 波改变的诊断与临床意义 [J]. 中国疗养医学, 2014, 23(8): 684-686. [5] Kania M, Fereniec M, Maniewski R. Evaluation of T-wave morphology dispersion in high-resolution ECG for risk stratification of sudden cardiac death [C]. //Computing in Cardiology. Krakow: IEEE, 2012: 785-788. [6] 王志毅,高克俭. 心电图形态诊断学 [M]. 天津:天津大学出版社, 2001: 299-300. [7] Boix M, Cantó B, Cuesta D, et al. Using the Wavelet Transform for T-wave alternans detection [J]. Mathematical and Computer Modelling, 2009, 50(5): 738-742. [8] Vila J., Gang Y., Presedo J., et al. A mathematical model for the T-wave pattern classification [C]. Computers in Cardiology 1998, 1998. [9] 王云鹤, 钱梦瑶, 刘定宇, 等. 心电信号多形态T波检测方法研究. 航天医学与医学工程, 2013, 26(4): 295-298. [10] Hadjem M, Nait-Abdesselam F. An ECG T-wave anomalies detection using a lightweight classification model for wireless body sensors [C]// International Conference on Communications Workshops. London: IEEE, 2015(6): 278-283. [11] 徐明芳. 基于规则的ECG信号ST-T段分类算法研究 [D]. 济南:山东大学, 2016. [12] Xu Mingfang, Wei Shoushui, Liu Chengyu, et al. Rule-based method for morphological classification of ST segment in ECG signals [J]. Journal of Medical and Biological Engineering, 2015, 35(6): 816-823. [13] Song Jinzhong, Yan Hong, Xiao Zhijun. Research on T-wave morphology analysis in ECG signal [C]//The 3rd International Conference on Biomedical Engineering and Informatics. Yantai: IEEE, 2010, 3: 1033-1036. [14] He Zhichao, Zhao Longzhang, Chen Chuang. Convolution neural network with multi-resolution feature fusion for facial expression recognition [J]. Laser and Optoelectronics Progress, 2018, 55(7): 370-375. [15] Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324. [16] Krizhevsky A, Sutskever I, Hinton G. ImageNet classification with deep convolutional neural networks [J]. Advances in Neural Information Processing Systems. 2012, 25(2): 1097-1105. [17] 刘明, 李国军, 郝华青, 等. 基于卷积神经网络的T波形态分类[J]. 自动化学报, 2016, 42(9): 1339-1346. [18] 董兵超, 于毅, 李振新. 基于MATLAB的心电信号的数字滤波处理 [J]. 数字技术与应用, 2012, 2012(10): 160-162. [19] 赵文哲, 方滨, 沈毅, 等. 心电信号中R波检测方法的比较研究 [J]. 生物医学工程学杂志, 2009, 26(1): 55-58. [20] Manriquez AI, Zhang Qinghua. An algorithm for QRS onset and offset detection in single lead electrocardiogram records [C] //Engineering in Medicine and Biology Society. Lyon: IEEE, 2007: 541-544. [21] Shang Haixia, Wei Shoushui, Liu Feifei, et al. An improved sliding window area method for T wave detection [J]. Comput Math Methods Med, 2019, 2019(1): 1-10. [22] Chon KH, Dash S, Ju K. Estimation of respiratory rate from photoplethysmogram data using time-frequency spectral estimation [J]. IEEE Trans Biomed Eng, 2009, 56(8): 2054-2063. [23] Yan Zhonghong, Miyamoto A, Jiang Zhongwei, et al. An overall theoretical description of frequency slice wavelet transform [J]. Mechanical Systems and Signal Processing, 2010, 24(2): 491-507. [24] Luo Kan, Li Jianqing, Wang Zhigang, et al. Patient-specific deep architectural model for ECG classification [J]. Journal of Healthcare Engineering, 2017, 2017(3): 1-13. [25] Xu Xiaoyan, Wei Shoushui, Ma Caiyun, et al. Atrial fibrillation beat identification using the combination of modified frequency slice wavelet transform and convolutional neural networks [J]. Journal of Healthcare Engineering, 2018, 2018(3): 1-8. [26] Chawla N, Bowyer KW, Hall LO, et al. SMOTE:Synthetic minority over-sampling technique [J]. Joumal of Artificial Intelligence Research, 2002, 16(1): 321-357. [27] Yan Zhonghong, Miyamoto A, Jiang Zhongwei, et al. An overall theoretical description of frequency slice wavelet transform [J]. Mechanical Systems and Signal Processing, 2009, 24(2): 491-507. [28] 郭网媚, 蔡宁. 卷积网络编码[J]. 中国电子科学研究院学报, 2012, 7(1): 10-16.
