An Image Classification Method Based on Adaptive Attention Mechanism and Feature Extraction Network

Table 1

Experimental results of Cifar-10 data set.

	AlexNet	VggNet	GoogLeNet	AA-ResNet	DenseNet
Train Acc	98.58%	98.93%	99.67%	99.96%	98.78%
Test Acc	91.42%	87.85%	90.54%	92.43%	91.11%
Loss	0.04569	0.03547	0.11288	0.00129	0.03543
Time (s)	145.71	202.68	306.90	253.71	374.37