Table 2 The results of different models on the RSSCN7 dataset.

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

Method	Pre	Rec	Acc	F1
ResNet34⁴⁸^†	0.8253	0.7588	0.7590	0.7450
ResNet50⁴⁸^†	0.7864	0.7147	0.7230	0.6861
ResNet101⁴⁸^†	0.8072	0.7228	0.7266	0.6793
AlexNet¹⁴^†	0.7938	0.7866	0.7878	0.7870
VGG11⁴⁶^†	0.8515	0.8470	0.8489	0.8452
VGG13⁴⁶^†	0.8393	0.8326	0.8237	0.8252
VGG16⁴⁶^†	0.8342	0.8216	0.8237	0.8224
VGG19⁴⁶^†	0.7583	0.7410	0.7374	0.7382
GoogleNet⁴⁷^†	0.8588	0.8602	0.8597	0.8582
Mobilenetv2⁵⁰^†	0.9027	0.9012	0.8993	0.9000
Mobilenetv3-l⁵⁰^†	0.8923	0.8937	0.8921	0.8906
Mobilenetv3-s⁵⁰^†	0.8421	0.8436	0.8417	0.8399
Shufflenet-0.5⁵¹^†	0.9345	0.9374	0.9353	0.9343
Shufflenet-1⁵¹^†	0.9632	0.9654	0.9640	0.9639
Shufflenet-1.5⁵¹^†	0.9410	0.9439	0.9424	0.9413
Shufflenet-2⁵¹^†	0.9674	0.9686	0.9676	0.9675
densenet121⁵²^†	0.9566	0.9570	0.9568	0.9565
densenet161⁵²^†	0.9713	0.9710	0.9712	0.9706
densenet169⁵²^†	0.9636	0.9641	0.9640	0.9635
densenet201⁵²^†	0.9557	0.9584	0.9568	0.9566
Efficient-b0-1k⁵³^†	0.9449	0.9451	0.9460	0.9444
Efficient-b1-1k⁵³^†	0.9396	0.9394	0.9388	0.9388
Efficient-b2-1k⁵³^†	0.9347	0.9350	0.9353	0.9343
Efficient-b3-1k⁵³^†	0.9633	0.9650	0.9640	0.9638
Efficient-b4-1k⁵³^†	0.9643	0.9650	0.9640	0.9643
Efficient-b5-1k⁵³^†	0.9400	0.9412	0.9388	0.9370
Efficient-b6-1k⁵³^†	0.9342	0.9380	0.9353	0.9347
Efficient-b7-1k⁵³^†	0.8786	0.8766	0.8777	0.8748
Efficientv2-l-1k⁵³^†	0.9310	0.9317	0.9317	0.9308
Efficientv2-m-1k⁵³^†	0.9242	0.9249	0.9245	0.9239
Efficientv2-s-1k⁵³^†	0.8908	0.8922	0.8921	0.8902
Convnext-s-1k⁵⁴^†	0.9403	0.9387	0.9388	0.9382
Convnext-b-22k⁵⁴^†	0.9438	0.9423	0.9424	0.9415
Convnext-b-1k⁵⁴^†	0.9527	0.9535	0.9532	0.9529
Convnext-t-1k⁵⁴^†	0.9665	0.9680	0.9676	0.9667
Convnext-l-1k⁵⁴^†	0.9450	0.9415	0.9424	0.9414
Convnext-l-22k⁵⁴^†	0.9608	0.9599	0.9604	0.9599
Convnext-xl-22k⁵⁴^†	0.9086	0.9065	0.9029	0.9021
DFAGCN⁵⁶^‡	–	–	0.9414	–
SNN-VGG-15⁴²^‡	–	–	0.9454	–
ViT-b-p16²⁴^†	0.9464	0.9486	0.9460	0.9471
ViT-b-p32²⁴^†	0.9089	0.9095	0.9101	0.9091
ViT-l-p16²⁴^†	0.9571	0.9579	0.9568	0.9574
TransResUNet⁶⁰^†	0.9654	0.9325	0.9338	0.9256
BPECN⁶¹^‡	–	–	0.9400	–
SKAL-AlexNet⁶²^‡	–	–	0.9335	–
SKAL-GoogleNet⁶²^‡	–	–	0.9575	–
SKAL-ResNet18⁶²^‡	–	–	0.9604	–
LFAGCU (ours)	0.9820	0.9820	0.9820	0.9818

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 2 The results of different models on the RSSCN7 dataset.

Search

Quick links