CardioGAN: Attentive Generative Adversarial Network with Dual Discriminators for Synthesis of ECG from PPG

Electrocardiogram (ECG) is the electrical measurement of cardiac activity, whereas Photoplethysmogram (PPG) is the optical measurement of volumetric changes in blood circulation. While both signals are used for heart rate monitoring, from a medical perspective, ECG is more useful as it carries additional cardiac information. Despite many attempts toward incorporating ECG sensing in smartwatches or similar wearable devices for continuous and reliable cardiac monitoring, PPG sensors are the main feasible sensing solution available. In order to tackle this problem, we propose CardioGAN, an adversarial model which takes PPG as input and generates ECG as output. The proposed network utilizes an attention-based generator to learn local salient features, as well as dual discriminators to preserve the integrity of generated data in both time and frequency domains. Our experiments show that the ECG generated by CardioGAN provides more reliable heart rate measurements compared to the original input PPG, reducing the error from 9.74 beats per minute (measured from the PPG) to 2.89 (measured from the generated ECG).

[1]  Susannah G. Fleming Lionel Tarassenko.  A Comparison of Signal Processing Techniques for the Extraction of Breathing Rate from the Photoplethysmogram , 2007 .

[2]  L. Nilsson,et al.  Respiration can be monitored by photoplethysmography with high sensitivity and specificity regardless of anaesthesia and ventilatory mode , 2005, Acta anaesthesiologica Scandinavica.

[3]  Trung Le,et al.  Dual Discriminator Generative Adversarial Nets , 2017, NIPS.

[4]  H. Huikuri,et al.  Time domain, geometrical and frequency domain analysis of cardiac vagal outflow: effects of various respiratory patterns. , 2001, Clinical physiology.

[5]  Xuan Zeng,et al.  HeartID: A Multiresolution Convolutional Neural Network for ECG-Based Biometric Human Identification in Smart Health Applications , 2017, IEEE Access.

[6]  Kira Radinsky,et al.  Improving ECG Classification Using Generative Adversarial Networks , 2020, AAAI.

[7]  Quan Liu,et al.  Electrocardiogram generation with a bidirectional LSTM-CNN generative adversarial network , 2019, Scientific Reports.

[8]  Zhen Yang,et al.  EmotionalGAN: Generating ECG to Enhance Emotion State Classification , 2019, Proceedings of the 2019 International Conference on Artificial Intelligence and Computer Science.

[9]  A. Etemad,et al.  Self-Supervised ECG Representation Learning for Emotion Recognition , 2020, IEEE Transactions on Affective Computing.

[10]  Kyle Ross,et al.  Classification of Cognitive Load and Expertise for Adaptive Simulation using Deep Multitask Learning , 2019, 2019 8th International Conference on Affective Computing and Intelligent Interaction (ACII).

[11]  Patrick E. McSharry,et al.  A dynamical model for generating synthetic electrocardiogram signals , 2003, IEEE Transactions on Biomedical Engineering.

[12]  Walter Karlen,et al.  Multiparameter Respiratory Rate Estimation From the Photoplethysmogram , 2013, IEEE Transactions on Biomedical Engineering.

[13]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[14]  David A. Clifton,et al.  Towards a Robust Estimation of Respiratory Rate from Pulse Oximeters , 2016 .

[15]  Kira Radinsky,et al.  PGANs: Personalized Generative Adversarial Networks for ECG Synthesis to Improve Patient-Specific Deep ECG Classification , 2019, AAAI.

[16]  Gari D Clifford,et al.  Synthetic ECG generation and Bayesian filtering using a Gaussian wave-based dynamical model , 2010, Physiological measurement.

[17]  拓海 杉山,et al.  “Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告 , 2017 .

[18]  Euan A. Ashley,et al.  Conquering the ECG , 2004 .

[19]  Nigel H. Lovell,et al.  The use of photoplethysmography for assessing hypertension , 2019, npj Digital Medicine.

[20]  Michael Muma,et al.  Computationally efficient heart rate estimation during physical exercise using photoplethysmographic signals , 2017, 2017 25th European Signal Processing Conference (EUSIPCO).

[21]  Philip H. S. Torr,et al.  Learn To Pay Attention , 2018, ICLR.

[22]  Helmut Alt,et al.  Computing the Fréchet distance between two polygonal curves , 1995, Int. J. Comput. Geom. Appl..

[23]  Kyle Ross,et al.  Toward Dynamically Adaptive Simulation: Multimodal Classification of User Expertise Using Wearable Devices , 2019, Sensors.

[24]  P. Hamilton,et al.  Open source ECG analysis , 2002, Computers in Cardiology.

[25]  Masoumeh Haghpanahi,et al.  Cardiologist-level arrhythmia detection and classification in ambulatory electrocardiograms using a deep neural network , 2019, Nature Medicine.

[26]  W. Kibbe,et al.  Investigating sources of inaccuracy in wearable optical heart rate sensors , 2020, npj Digital Medicine.

[27]  J. Vagedes,et al.  How accurate is pulse rate variability as an estimate of heart rate variability? A review on studies comparing photoplethysmographic technology with an electrocardiogram. , 2013, International journal of cardiology.

[28]  P. Laguna,et al.  Photoplethysmography pulse rate variability as a surrogate measurement of heart rate variability during non-stationary conditions , 2010, Physiological measurement.

[29]  Ali Etemad,et al.  Self-Supervised Learning for ECG-Based Emotion Recognition , 2020, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[30]  A. Awad,et al.  The Use of Joint Time Frequency Analysis to Quantify the Effect of Ventilation on the Pulse Oximeter Waveform , 2006, Journal of Clinical Monitoring and Computing.

[31]  Jan C. Brammer,et al.  NeuroKit2: A Python toolbox for neurophysiological signal processing. , 2021, Behavior research methods.

[32]  Min Wu,et al.  Learning Your Heart Actions From Pulse: ECG Waveform Reconstruction From PPG , 2019, bioRxiv.

[33]  Loïc Le Folgoc,et al.  Attention U-Net: Learning Where to Look for the Pancreas , 2018, ArXiv.

[34]  Kristof Van Laerhoven,et al.  Introducing WESAD, a Multimodal Dataset for Wearable Stress and Affect Detection , 2018, ICMI.

[35]  G. Burch.  [Cardiovascular diseases]. , 1956, Revista medica cubana.

[36]  Kristof Van Laerhoven,et al.  Deep PPG: Large-Scale Heart Rate Estimation with Convolutional Neural Networks , 2019, Sensors.

[37]  Derek Abbott,et al.  Systolic Peak Detection in Acceleration Photoplethysmograms Measured from Emergency Responders in Tropical Conditions , 2013, PloS one.