Table 3 The results of different models on the WHU-RS19 dataset.

From: Local feature acquisition and global context understanding network for very high-resolution land cover classification

Method	Pre	Rec	Acc	F1
AlexNet¹⁴^†	0.6031	0.5908	0.5833	0.5591
VGG11⁴⁶^†	0.8024	0.7904	0.7708	0.7666
VGG13⁴⁶^†	0.7955	0.8096	0.7812	0.7781
VGG16⁴⁶^†	0.8704	0.8509	0.8438	0.8446
VGG19⁴⁶^†	0.8341	0.8412	0.8333	0.8184
GoogleNet⁴⁷^†	0.8518	0.8228	0.8125	0.8153
ResNet34⁴⁸^†	0.9089	0.8421	0.8333	0.8371
ResNet50⁴⁸^†	0.8956	0.8509	0.8646	0.8395
ResNet101⁴⁸^†	0.8614	0.7728	0.7604	0.7578
Mobilenetv2⁵⁰^†	0.9895	0.9829	0.9792	0.9848
Mobilenetv3-l⁵⁰^†	0.9722	0.9671	0.9688	0.9661
Mobilenetv3-s⁵⁰^†	0.9912	0.9868	0.9896	0.9877
Shufflenet- × 0.5⁵¹^†	0.9529	0.9566	0.9479	0.9505
Shufflenet- × 1⁵¹^†	0.9868	0.9912	0.9896	0.9877
Shufflenet- × 1.5⁵¹^†	0.9836	0.9803	0.9792	0.9800
Shufflenetv2- × 2⁵¹^†	0.9912	0.9868	0.9896	0.9877
densenet121⁵²^†	0.9895	0.9868	0.9896	0.9866
densenet161⁵²^†	0.9895	0.9868	0.9896	0.9866
densenet169⁵²^†	0.9912	0.9868	0.9896	0.9877
densenet201⁵²^†	0.9868	0.9868	0.9896	0.9850
Efficient-b0-1k⁵³^†	0.9675	0.9715	0.9688	0.9660
Efficient-b1-1k⁵³^†	0.9781	0.9763	0.9792	0.9743
Efficient-b2-1k⁵³^†	0.9866	0.9789	0.9792	0.9797
Efficient-b3-1k⁵³^†	0.9722	0.9741	0.9688	0.9705
Efficient-b4-1k⁵³^†	0.9942	0.9934	0.9896	0.9934
Efficient-b5-1k⁵³^†	0.9868	0.9868	0.9896	0.9850
Efficient-b6-1k⁵³^†	0.9942	0.9884	0.9879	0.9913
Efficient-b7-1k⁵³^†	0.9538	0.9465	0.9479	0.9465
Efficientv2-l-1k⁵³^†	0.9825	0.9759	0.9688	0.9771
Efficientv2-m-1k⁵³^†	0.9876	0.9846	0.9792	0.9855
Efficientv2-s-1k⁵³^†	0.9412	0.9368	0.9271	0.9284
Convnext-s-1k⁵⁴^†	0.9749	0.9697	0.9688	0.9694
Convnext-b-22k⁵⁴^†	0.9688	0.9658	0.9688	0.9634
Convnext-b-1k⁵⁴^†	0.9925	0.9895	0.9896	0.9901
Convnext-t-1k⁵⁴^†	0.9722	0.9697	0.9688	0.9677
Convnext-l-1k⁵⁴^†	0.9810	0.9803	0.9792	0.9784
Convnext-l-22k⁵⁴^†	0.9732	0.9803	0.9688	0.9743
Convnext-xl-22k⁵⁴^†	0.9807	0.9737	0.9792	0.9743
MLLD⁵⁷^‡	–	–	0.9069	–
HSL-MINet⁵⁸^‡	–	–	0.9122	–
ViT-b-p16²⁴^†	0.9810	0.9803	0.9792	0.9784
ViT-b-p32²⁴^†	0.9722	0.9671	0.9688	0.9661
ViT-l-p16²⁴^†	0.9912	0.9868	0.9896	0.9877
TransResUNet⁶⁰^†	0.9656	0.9566	0.9583	0.9568
BPECN⁶¹^‡	–	–	0.9820	–
SF-MSFormer-ResNet18⁶⁴‡	–	–	0.9860	–
LGFormer	–	–	0.9920	–
LFAGCU (ours)	0.9942	0.9934	0.9896	0.9934

Bold indicates the optimal solution, ^‡represents the data results from the reference literature, and ^†represents the experimental results based on the settings of this paper's parameters.

Back to article page

Table 3 The results of different models on the WHU-RS19 dataset.

Search

Quick links