Table 1.

Comparison of recall, precision, F_β score, validation accuracy and training time for different models.

Dataset	Structural design	Recall	Precision	Fβ (0.5) score	Validation Accuracy	No. parameters	Training time (seconds)
Sample Dataset	Vanilla gray	0.50	0.58	0.56	50.7%	321225	2
	Vanilla RGB	0.59	0.62	0.61	51.8%	322793	2
	Hybrid CNN VGG	0.56	0.65	0.63	68%	15252133	16
	VDSNet	0.64	0.62	0.64	70.8%	15488051	19
	Modified CapsNet	0.42	0.71	0.45	59%	12167424	37
	Basic CapsNet	0.60	0.62	0.62	57%	14788864	75
Full Dataset	Vanilla gray	0.58	0.68	0.66	67.8%	321225	51
	Vanilla RGB	0.61	0.68	0.66	69%	322793	53
	Hybrid CNN VGG	0.62	0.68	0.67	69.5%	15252133	384
	VDSNet	0.63	0.69	0.68	73%	15488051	431
	Modified CapsNet	0.48	0.61	0.58	63.8%	12167424	856
	Basic CapsNet	0.51	0.64	0.61	60.5%	14788864	1815