Table 8. Comparison of the proposed ECOVNet with other state-of-the-art methods on COVID-19 detection.

Method	Precision (%) (COVID-19)	Recall (%) (COVID-19)	Accuracy (%)
COVID-Net (Wang, Lin & Wong, 2020)	92.80	91.00	93.92
EfficientNet-B3 (Luz et al., 2020)	95.29	81.00	94.49
DarkCovidNet (Ozturk et al., 2020)	96.00	88.00	92.00
CoroNet (Khan, Shah & Bhat, 2020)	91.00	87.00	88.00
ECOVNet-Hard Ensemble (Proposed)	94.17	97.00	96.07
ECOVNet-Soft Ensemble (Proposed)	92.59	100	96.07

Note:

Bold indicates that the method has statistically better performance than other methods.