Fig 1.
Model establishment of the human heart.
(A) 3D geometry. (B) Finite element model. (C) Initial condition.
Fig 2.
Action potential of sample points in model 0.
Fig 3.
Exponentially attenuated sinusoidal function.
Fig 4.
Action potential of sample points in model 1.
Fig 5.
Action potential of sample points in model 2.
Fig 6.
Different patterns of electrocardio signals with Gaussian noise.
X-axis represents time, Y-axis represents the magnitude of the membrane potential.
Fig 7.
Comparison of different optimization algorithms.
(A) Schematic diagram of test function. (B) Convergence comparison.
Fig 8.
Kurtograms of different signals.
(A) Kurtogram of signal 0. (B) Kurtogram of signal 1. (C) Kurtogram of signal 2. X-axis represents the frequency bands, Y-axis represents the resolution levels, color intensity represents the kurtosis value (higher intensity indicates higher kurtosis). Kurtograms exhibit similar impulses components but differentiated characteristics of the kurtosis values across three signals.
Fig 9.
Sensitivity analysis of parameter K and in VMD.
(A) Influence of K on the VMD performance. (B) Influence of on the VMD performance. X-axis represents the value of K or
, left Y-axis represents spectral kurtosis, right Y-axis represents KL divergence. The parameters K and
have significant and coupled impacts on decomposition performance.
Fig 10.
Pareto optimal front result of MOCOA.
Table 1.
Summary of key parameters and performance indices of VMD.
Fig 11.
Comparison of original VMD and MOCOA-VMD results.
(A) Original VMD result. (B) Optimal VMD result based on MOCOA. IMF1’s MOCOA-VMD result demonstrates superior performance in capturing the main component of the signal compared to the original VMD.
Fig 12.
Deep attention network architecture for arrhythmia classification.
Table 2.
Comparison of ablation study. The deep model which consists of MOCOA-VMD and attention scheme achieves the best accuracy.
Fig 13.
Bayesian optimization results of key hyperparameters.
Fig 14.
Confusion matrix.
Fig 15.
Performance validation on the MIT-BIH database.
(A) Training state. (B) Confusion matrix.
Table 3.
Performance comparison with other papers.