Table 1.

Summary of the main related work. The last column indicates whether the learning is of supervised (S) or unsupervised (U) type. See also [39] and Rahman et al. (2021) for a review on the diagnosis of COVID-19 from radiography images.

Family	Work	Approach	Type
X-ray	[28]	Multiresolution	S
	[29]	Deep learning	S
	[14]	Majority voting	S
	[51]	U-Net + Transfer Learning	S
	[19]	Shallow ML	S
CT segmentation	[55]	NormNet	U
	[41]	U-Net + SegNet	S
	[18]	InfNet	S
CT binary	[43]	CNN (CTnet10)	S
classification	[50]	VGG-16	S
	[17]	WDNN	S
	[38]	Fusion + Majority voting	S
	[47]	Fusion + Majority voting	S
Autoencoders	[54]	SSAE + Softmax	S
	[15]	CAE + Softmax	S
	[36]	SSAE + Softmax	S
	Our approach	DDCAE + hidden features	U