Table 14.

Performance of similar segmentation-classification pipelines without dataset composition issues.

Research paper	Seg. DSC	Acc.	COV. Sens.
Yeh et al. [20]
3-class	0.88	–	82%
Wehbe et al. [21]
2-class	–	82%	75%
Abdullah et al. [25]
2-class	0.96	79%	–
Wang et al. [26]
2-class	0.89	91%	92%
Tabik et al. [22]
2-class (COVID-GR-1.0 dataset)	0.885	76%	73%
Ours
Best 3-class Ens. (Maj. Vot.)	0.95	84%	82%
Best 2-class Ens. (Wei. Avg.)	0.95	91%	92%
2-class (COVID-GR-1.0 dataset)	0.95	77%	78%