Table 7 Assessing the efficacy of the proposed model against conventional DL models using the test dataset. The best-achieved values are in bold, while the second-highest values are underlined. The top values are individually highlighted in bold and underlined for both normal and abnormal categories. “N” and “A,” represent normal and abnormal labels. [Values in %].

From: An interpretable framework for gastric cancer classification using multi-channel attention mechanisms and transfer learning approach on histopathology images

	Model	Class	Sens	Spec	F1	Pre	Avg. Acc
VT	ViT⁹²	N	74.08	78.90	75.82	74.85
	ViT⁹²	A	74.23	77.90	75.89	75.01	77.86
	CaiT⁹³	N	75.68	72.44	75.74	73.25
	CaiT⁹³	A	73.44	75.68	73.73	73.31	74.80
	DeiT⁹⁴	N	93.22	94.12	93.85	93.88
	DeiT⁹⁴	A	92.36	94.77	93.68	93.46	93.87
	CoaT⁹⁶	N	73.89	81.25	80.25	80.22
	CoaT⁹⁶	A	87.58	73.89	82.32	83.14	80.92
	BoTNet-50⁹⁸	N	95.24	94.28	95.20	95.40
	BoTNet-50⁹⁸	A	93.25	94.52	95.65	93.82	95.03
	LeViT⁹⁷	N	79.62	80.23	81.24	80.28
	LeViT⁹⁷	A	81.25	79.25	80.41	80.40	80.74
	T2T-ViT⁹⁵	N	90.14	92.35	91.02	88.98
	T2T-ViT⁹⁵	A	89.35	91.25	92.23	90.20	92.01
CNN	VGG-16⁸⁴	N	96.72	98.50	97.72	95.68
	VGG-16⁸⁴	A	98.45	96.72	97.69	95.66	97.72
	MCLNet¹⁴⁶	N	97.57	97.03	97.77	97.78
	MCLNet¹⁴⁶	A	97.03	97.57	96.67	95.66	97.36
	Xception⁸⁶	N	97.85	98.47	98.23	98.32
	Xception⁸⁶	A	98.55	98.71	98.24	97.42	97.98
	Inception-V3⁸⁷	N	98.13	98.38	98.26	97.66
	Inception-V3⁸⁷	A	98.34	98.13	98.23	97.66	98.01
	AlexNet¹³⁶	N	96.42	95.32	94.21	92.36
	AlexNet¹³⁶	A	92.57	96.74	94.99	94.44	94.90
	DenseNet-121⁸⁵	N	97.12	97.76	95.09	95.22
	DenseNet-121⁸⁵	A	97.67	96.16	96.90	95.36	96.92
	MobileNetV3⁵⁶	N	96.57	92.83	93.87	92.66
	MobileNetV3⁵⁶	A	92.83	96.57	90.80	90.88	92.64
	InceptionResNet-V1⁸⁹	N	95.47	96.74	96.40	95.78
	InceptionResNet-V1⁸⁹	A	95.55	95.23	95.32	94.22	96.12
	DenseNet-121⁵⁶	N	96.57	92.83	93.87	94.82
	DenseNet-121⁵⁶	A	92.83	96.57	90.80	89.66	92.64
	ResNet-50⁸³	N	93.41	93.22	96.55	92.64
	ResNet-50⁸³	A	95.77	93.22	94.62	94.44	94.67
	EfficientNetV2¹⁴⁷	N	98.37	97.60	98.40	98.42
	EfficientNetV2¹⁴⁷	A	97.60	98.37	97.53	97.12	98.06
MLP	gMLP¹⁴³	N	89.23	88.32	89.60	87.56
	gMLP¹⁴³	A	88.69	88.70	88.24	87.66	88.51
	MLP-Mixer¹⁴⁴	N	72.11	73.01	73.32	72.22
	MLP-Mixer¹⁴⁴	A	72.41	74.21	72.02	73.58	73.24
	ResMLP¹⁴⁵	N	72.12	78.41	75.21	74.22
	ResMLP¹⁴⁵	A	73.21	77.14	75.21	73.24	74.73
Proposed	MCAM	N	98.76	98.53	98.72	98.50
Proposed	MCAM	A	98.53	98.76	98.50	97.87	98.48

Back to article page

Search

Quick links