CAW: A Remote-Sensing Scene Classification Network Aided by Local Window Attention

Table 7

Overall accuracy and other parameters of the method on the WHU-RS19 dataset.

Model	Accuracy (%)	Precision (%)	Recall (%)	Specificity (%)	F1-score (%)
ResNet50 [31]	94.66	95.15	94.62	99.71	94.62
Vgg16 [32]	94.66	95.44	94.61	99.71	94.78
GoogleNet [27]	90.29	90.89	90.50	99.47	90.27
DenseNet [33]	95.15	96.08	95.14	99.73	95.34
ViT-Ti [14]	82.04	83.74	82.11	99.01	82.28
Swin transformer [20]	91.26	92.25	91.35	99.52	91.25
VAN [17]	93.67	94.44	93.72	99.66	93.59
Conformer-Ti [15]	95.63	95.75	95.54	99.76	95.55
CMT [26]	95.63	96.18	95.68	99.76	95.77
MPViT [25]	95.63	95.93	95.75	99.76	95.69
CAW	96.12	96.23	96.03	99.79	95.96