Table 9.

Performance of other recent works on the Kermany et al. [10] dataset with values rounded off to the nearest two decimal positions

Authors	Classes	Technique	Accuracy (%)	Precision (%)	Recall (%)	AUC (%)
Kermany et al. [10]	Normal and Pneumonia	Inception V3 pretrained CNN model	92.8	90.1	93.2	–
Nahida et al. [27]	Normal and Pneumonia	Two-channel CNN model	97.92	98.38	97.47	97.97
Stephen et al. [30]	Normal and Pneumonia	Custom CNN model without Transfer Learning	93.73	–	–	–
Chouhan et al. [14]	Normal and Pneumonia	Majority voting ensemble model	96.39	93.28	99.62	99.34
Rajaraman et al. [47]	Normal and Pneumonia	Custom VGG-16 model	96.2	97.0	99.5	99.0
Siddiqi et al. [19]	Normal and Pneumonia	Deep sequential CNN model	94.39	92.0	99.0	–
Hashmi et al. [48]	Normal and Pneumonia	Weighted classifier	98.43	–	–	99.76
Yu Xiang et al. [33]	Normal and Pneumonia	CGNET	98.72	97.48	99.15	–
El Asnaoui et al. [22]	Normal and Pneumonia	Deep CNN model	96.27	98.06	94.61	–
Saraiva et al. [16]	Normal and Pneumonia	MLP and NN approach	92.16	–	–	–
Saraiva et al. [17]	Normal and Pneumonia	Custom CNN	95.30	–	–	–
Mittal et al. [34]	Normal and Pneumonia	CapsNet architecture	96.36	–	–	–
Rahman et al. [21]	Normal and Pneumonia	Deep CNN model	98.0	97.0	99.0	98.0
Sagar Kora Venu et al. [5]	Normal and Pneumonia	Weighted average ensemble model	98.46	98.38	99.53	99.60
Toğaçar et al. [49]	Normal and Pneumonia	Deep CNN model	96.84	96.88	96.83	96.80
Nahida et al. [25]	Normal and Pneumonia	SMOTE on ensembled features from VGG-19 and CheXNet	98.90	–	–	99.00
Islam et al. [28]	Normal and Pneumonia	Feature concatenations with ANN	98.99	99.18	98.90	–
Proposed Work	Normal and Pneumonia	Stacking classifier based on features extracted from Xception	98.3	99.29	98.36	98.24