Table 4.
Overview of related works based on various input types.
| Refs. | Method | Input data type |
Detection task | Performance % |
|---|---|---|---|---|
| 31 |
Time–frequency analysis of PCG signal using chirplet transform |
PCG | Valve disease diagnosis | Accuracy 98.33 |
| 32 | Recurrent neural network with long short-term memory | CCTA | Calcified plaque detection |
Accuracy 90.3 Sensitivity 92.1 Specificity 88.9 |
| 33 | CNN | ECG | Diagnosis of different cardiovascular diseases | Accuracy 95 |
| 34 | Optimal time–frequency concentrated biorthogonal wavelet-based features | ECG | CAD diagnosis | Accuracy 98.53 |
| 35 | Binomial rendition of the bivariate mixed-effects regression model |
CCTA, ECG |
CAD diagnosis |
Sensitivity 99 Specificity 88 |
| 36 | Discrete wavelet transform, multivariate multi-scale entropy, | ECG | CAD diagnosis | Accuracy 98.67 |
| 37 | Sequential minimal optimization, Naive Bayes, and ensemble algorithm | ECG | CAD diagnosis | Accuracy 88.5 |
| 38 | Computing complex ventricular excitation index | Magneto-cardiography | CAD diagnosis |
Sensitivity 91 Specificity 84 |
| 39 | Extracted time- and frequency-domain features from PCG signal as inputs to neural network classifier | PCG | CAD diagnosis |
Accuracy 82.57 Sensitivity 85.61 Specificity 79.55 |
| 40 | Multimodal feature fusion and hybrid feature selection, SVM classifier |
ECG, PCG |
CAD diagnosis |
Accuracy 96.67 Sensitivity 96.67 Specificity 96.67 F1-measure 96.64 |
| 41 | Multimodal feature fusion, SVM classifier |
PCG, PPG |
CAD diagnosis |
Sensitivity 80 Specificity 93 |
| 42 |
Combined feature set related to heart rate variability and shape of PPG waveform, SVM classifier Two sets of features extracted from |
PPG | CAD diagnosis |
Sensitivity 73 Specificity 87 |
| 43 | Two sets of features extracted from PPG and PCG, SVM classifier |
PCG, PPG |
CAD diagnosis |
Sensitivity 92 Specificity 90 |
| 44 | Novel feature representation using synchrosqueezing transform, CAD diagnosis based on entropy of PCG, SVM classifier | PCG | CAD diagnosis | Accuracy 83.48 |
| 45 | Hybrid neural network-genetic algorithm | Echo | CAD diagnosis |
Accuracy 93.85 Sensitivity 97 Specificity 92 |
| 46 |
Sequential minimal optimization Naive Bayes, C4.5 and AdaBoost |
Laboratory data, echo |
CAD diagnosis | Accuracy 82% |
| 19 | Rotation forest with neural networks as base classifiers | Cleveland | CAD diagnosis |
Accuracy 91.20 AUC 91.50 Sensitivity 95.60 Specificity 86.70 |
| 21 | Nested ensemble nu-Support Vector Classification | Z-Alizadeh Sani | CAD diagnosis |
Accuracy 94.66 Precision 94.70 Sensitivity 94.70 |
| 22 | Ensemble PSO-based fuzzy rule extraction | Cleveland | CAD diagnosis |
Accuracy 92.59 Specificity 94.37 Sensitivity 90.51 |
| Proposed method | Random forest, CNNs as feature extractors, Adam optimizer | CMR | CAD diagnosis |
Accuracy 99.18 Sensitivity 98.88 Specificity 99.66 AUC 99 |
ECG Electrocardiograph, Echo Echocardiography, PCG Phonocardiograph, PPG Photoplethysmography, SVM Support vector machine.