Automobile Component Recognition Based on Deep Learning Network with Coarse-Fine-Grained Feature Fusion

Table 3

Performance comparison of SOTA models.

Model	Training	Testing		Evaluation		Evaluation with Algorithm 1
	Dataset	Loss	Acc. (%)	Loss	Acc. (%)	Loss	Acc. (%)
BiT	DATA_ORI	0.1001	92.15	2.6926	30.23	4.9752	33.51
	DATA_CON	0.2318	90.84	3.3496	34.59	2.4041	47.03
CvT	DATA_ORI	1.0438	83.33	4.7235	24.05	3.4118	44.05
	DATA_CON	0.9596	84.96	3.2217	33.24	2.9417	49.19
ViT	DATA_ORI	3.4291	77.43	13.7787	31.26	9.5347	36.54
	DATA_CON	2.1206	85.07	12.8839	34.37	8.5453	40.62
PDLN	DATA_ORI	0.1303	94.38	2.7066	50.86	3.1891	54.05
	DATA_CON	0.1945	95.11	1.7629	61.62	0.2841	92.43