CausalCervixNet: convolutional neural networks with causal insight (CICNN) in cervical cancer cell classification—leveraging deep learning models for enhanced diagnostic accuracy

Table 11 The performance analysis of the proposed CausalCervixNet method

Dataset	Models	Avg. Precision	Avg. Recall	Avg. F1-score	Avg. Accuracy	AUC	Time (sec)
SIPaKMeD	CICNN-VGG16	0.988	0.987	0.988	98.76%	0.992	10,000
	CICNN-VGG19	0.986	0.985	0.985	98.52%	0.991	10,500
	CICNN-ResNet-50	0.991	0.991	0.991	99.14%	0.995	11,200
	CICNN-XceptionNet	0.881	0.861	0.862	86.21%	0.914	11,800
ShUCSEIT	CICNN-VGG16	0.971	0.971	0.971	97.05%	0.982	6300
	CICNN-VGG19	0.969	0.970	0.970	96.82%	0.980	6700
	CICNN-ResNet-50	0.991	0.991	0.991	99.09%	0.994	7400
	CICNN-XceptionNet	0.933	0.933	0.932	93.18%	0.957	8000
Herlev	CICNN-VGG16	0.898	0.860	0.871	86.02%	0.918	3600
	CICNN-VGG19	0.876	0.853	0.859	84.95%	0.912	3900
	CICNN-ResNet-50	0.979	0.978	0.978	97.31%	0.984	4500
	CICNN-XceptionNet	0.837	0.786	0.793	79.57%	0.881	5000

ISSN: 1471-2407