Rede Neural IncResU-Net para Inferência de Sinais Eletrocardiograma a partir de Sinais Fotopletismograma
Abstract
The electrocardiogram (ECG) is a medical exam that measures the electrical activity of the heart, while the photoplethysmogram (PPG) measures changes in blood volume through light-based technology. Although both methods are used to monitor heart rate, ECG is considered the gold standard method for diagnosing heart disease because it provides additional information about heart function. Despite efforts to integrate ECG detection into wearable devices for continuous and reliable heart monitoring, PPG sensors are currently the main viable solution. This work proposes a method called PPG2ECG, based on domain mapping, applying a set of convolution filters, using a U-net Inception neural network architecture to infer ECG signals from PPG signals. To evaluate the effectiveness of the proposed method, two evaluation strategies were adopted, based on personalized and generalized models. The results obtained demonstrated average error values of 0.015 and 0.026, respectively.
References
Berkaya, S. K. et al. (2018). A survey on ecg analysis. Biomedical Signal Processing and Control, 43:216–235.
Chiu, H.-Y., Shuai, H.-H., and Chao, P.-C. (2020). Reconstructing qrs complex from ppg by transformed attentional neural networks. IEEE Sensors Journal, 20(10):12374–12383.
Elgendi, M. et al. (2019). The use of photoplethysmography for assessing hypertension. npj Digital Medicine, 6(Volume 2).
Karlen, W., Raman, S., Ansermino, J. M., and Dumont, G. A. (2013). Multiparameter respiratory rate estimation from the photoplethysmogram. IEEE transactions on biomedical engineering, 60(7):1946–1953.
Liang, Y., Chen, Z., Ward, R., and Elgendi, M. (2018). Hypertension assessment via ecg and ppg signals: An evaluation using mimic database. Diagnostics, 8(3):65.
Organization, W. H. (2021). World health statistics 2020: monitoring health for the SDGs, sustainable development goals. World Health Organization, Geneve, Switzerland.
Pimentel, M. A., Clifton, D. A., and Clifton, L. (2017). Toward a robust estimation of respiratory rate from pulse oximeters. IEEE Transactions on Biomedical Engineering, 64(8):1914–1923.
Ravichandran, V., Murugesan, B., Shankaranarayana, S. M., Ram, K., S.P., P., Joseph, J., and Sivaprakasam, M. (2019). Deep network for capacitive ecg denoising. In 2019 IEEE International Symposium on Medical Measurements and Applications (MeMeA), pages 1–6.
Reiss, A., Indlekofer, I., Schmidt, P., and Van Laerhoven, K. (2019). Deep ppg: Large-scale heart rate estimation with convolutional neural networks. Sensors, 19(14):3079.
Ronneberger, O., Fischer, P., and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Lecture Notes in Computer Science, pages 234–241. Springer International Publishing.
Sarkar, P. and Etemad, A. (2020). Cardiogan: Attentive generative adversarial network with dual discriminators for synthesis of ecg from ppg. In AAAI Conference on Artificial Intelligence.
Schmidt, P., Reiss, A., Duerichen, R., Marberger, C., and Van Laerhoven, K. (2018). Introducing wesad, a multimodal dataset for wearable stress and affect detection. In Proceedings of the 20th ACM International Conference on Multimodal Interaction, pages 400–408. ACM.
Szegedy, C., Ioffe, S., Vanhoucke, V., and Alemi, A. A. (2017). Inception-v4, inceptionresnet and the impact of residual connections on learning. In Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence, AAAI’17, page 4278–4284. AAAI Press.
Tian, X., Zhu, Q., Li, Y., and Wu, M. (2020). Cross-domain joint dictionary learning for ecg reconstruction from ppg. In IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP).
Zhu, Q., Tian, X., Wong, C.-W., and Wu, M. (2019). Ecg reconstruction via ppg: A pilot study. In 2019 IEEE EMBS International Conference on Biomedical Health Informatics (BHI), pages 1–4.
