Using Machine Learning Algorithms to Determine the Post-COVID State of a Person by Their Rhythmogram

Figures

• 

  Figure 1. Illustration depicting the setup for recording and analyzing the cardiac rhythmogram system.

• 

  Figure 2. Dilated convolution with kernel = 3 and its receptive field.

• 

  Figure 3. The residual block.

• 

  Figure 4. Spike detector circuit.

• 

  Figure 5. Head block.

• 

  Figure 6. Segment of the rhythmogram of a patient with COVID-19 comprising 400 counts, displaying 9 instances of spike anomalies. These anomalies exhibit a distinct repetitive pattern and varying amplitudes within the range of ±100 ms from the mean value.

• 

  Figure 7. An illustration of a collection of normalized spikes from diverse subjects who have been diagnosed with COVID-19.

• 

  Figure 8. Comparing rhythmograms obtained from various devices in a patient with COVID-19 over a long-term period: (a) Poly-Spectrum-8; (b) Zephyr.

• 

  Figure 9. Comparison of rhythmograms in a patient with COVID-19 obtained using different algorithms: (a) Pan–Tompkins; (b) Hamilton; (c) Two-Moving-Average.

• 

  Figure 10. The space of parameters from the biochemistry module of indicators reflecting the state of the blood.

• 

  Figure 11. The space of parameters from the hematological studies module of indicators reflecting the state of the blood.

• 

  Figure 12. The space of parameters from the hemostasis module of indicators reflecting the state of the blood.