Junsang Park1, Junho An2, Jinkook Kim3, Sunghoon Jung4, Yeongjoon Gil5, Yoojin Jang6, Kwanglo Lee7, Il-Young Oh8. 1. HUINNO Co., Ltd., Seoul, Republic of Korea. 2. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: junho@huinno.com. 3. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: jinkook@huinno.com. 4. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: shjung@huinno.com. 5. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: kyzoon@huinno.com. 6. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: yjluca98@huinno.com. 7. HUINNO Co., Ltd., Seoul, Republic of Korea. Electronic address: kwanglo@huinno.com. 8. Department of Internal Medicine, Seoul National University, Seoul National University Bundang Hospital, Seongnam, Republic of Korea. Electronic address: spy510@snu.ac.kr.
Abstract
BACKGROUND AND OBJECTIVES: Most deep-learning-related methodologies for electrocardiogram (ECG) classification are focused on finding an optimal deep-learning architecture to improve classification performance. However, in this study, we proposed a methodology for fusion of various single-lead ECG data as training data in the single-lead ECG classification problem. METHODS: We used a squeeze-and-excitation residual network (SE-ResNet) with 152 layers as the baseline model. We compared the performance of a 152-layer SE-ResNet trained on ECG signals from various leads of a standard 12-lead ECG system to that of a 152-layer SE-ResNet trained on only single-lead ECG data with the same lead information as the test set. The experiments were performed using five different types of rhythm-type single-lead ECG data obtained from Konkuk University Hospital in South Korea. RESULTS: Experiment results based on the combination from the relationship experiments of the leads showed that lead -aVR or II revealed the best classification performance. In case of -aVR, this model achieved a high F1 score for normal (98.7%), AF (98.2%), APC (95.1%), and VPC (97.4%), indicating its potential for practical use in the medical field. CONCLUSION: We concluded that the 152-layer SE-ResNet trained by fusion of single-lead ECGs had better classification performance than the 152-layer SE-ResNet trained on only single-lead ECG data, regardless of the single-lead ECG signal type. We also found that the best performance directions for single-lead ECG classification are Lead -aVR and II.
BACKGROUND AND OBJECTIVES: Most deep-learning-related methodologies for electrocardiogram (ECG) classification are focused on finding an optimal deep-learning architecture to improve classification performance. However, in this study, we proposed a methodology for fusion of various single-lead ECG data as training data in the single-lead ECG classification problem. METHODS: We used a squeeze-and-excitation residual network (SE-ResNet) with 152 layers as the baseline model. We compared the performance of a 152-layer SE-ResNet trained on ECG signals from various leads of a standard 12-lead ECG system to that of a 152-layer SE-ResNet trained on only single-lead ECG data with the same lead information as the test set. The experiments were performed using five different types of rhythm-type single-lead ECG data obtained from Konkuk University Hospital in South Korea. RESULTS: Experiment results based on the combination from the relationship experiments of the leads showed that lead -aVR or II revealed the best classification performance. In case of -aVR, this model achieved a high F1 score for normal (98.7%), AF (98.2%), APC (95.1%), and VPC (97.4%), indicating its potential for practical use in the medical field. CONCLUSION: We concluded that the 152-layer SE-ResNet trained by fusion of single-lead ECGs had better classification performance than the 152-layer SE-ResNet trained on only single-lead ECG data, regardless of the single-lead ECG signal type. We also found that the best performance directions for single-lead ECG classification are Lead -aVR and II.
Authors: Georgios Petmezas; Leandros Stefanopoulos; Vassilis Kilintzis; Andreas Tzavelis; John A Rogers; Aggelos K Katsaggelos; Nicos Maglaveras Journal: JMIR Med Inform Date: 2022-08-15