Table 2.

Pseudocode of the optimal classifier selection algorithm.

Step 1:	load the COVID-19 dataset.
Step 2:	load the ResNet-18.
Step 3:	modify the structure of the ResNet-18 based on the labels of the COVID-19 dataset using ELM and RVFL, respectively.
Step 4:	train the two models and test them based on 5-fold cross-validation.
Step 5:	compute the average testing accuracies of the two models.
Step 6:	output the CGENet with the better classifier, which yielded the highest average testing accuracy.