Table 9.

Performance of past DenseNet-Based models versus radiologists.

Paper Reviewed	F1	ACC	COVID-19 Sens.
Yeh et al. [41]
3-class	–	–	81.82%
Haghanifar et al. [11]
2-class	94%	98.62%	–
3-class	85%	81.04%	–
Mangal et al. [23]
3-class	92.3%	90.5%	100%
Al-Waisy et al. [3]
2-class	99.99%	99.99%	99.98%
Rajaraman et al. [30]
4-class	96.77%	96.83%	96.34%
Radiologists [39]
2-class	–	81%	70%

Note: Haghanifar et al. [11], Mangal et al. [23], Al-Waisy et al. [3], and Rajaraman et al. [30] all improperly used Kermany et al.’s dataset [18].