A Deep Learning Method for Building Extraction from Remote Sensing Images by Fuzing Local and Global Features

Table 1

Evaluation indicators of different networks.

Method	P (%)	R (%)	F1 (%)	IOU (%)	Acc (%)
Segnet	95.01	74.82	83.71	81.88	93.20
Unet	93.71	81.93	87.42	85.43	94.51
MAP-Net	93.79	90.82	92.28	89.40	96.17
MSRF-Net (k = 38)	94.97	91.68	93.29	90.16	96.65
TransUnet	93.75	87.21	90.36	88.49	95.66
TransFuse-L	94.17	90.18	92.12	89.93	96.84
MATUnet	95.05	92.23	93.62	92.14	97.06

Note. Bold numbers indicate the best performance of each indicator.