Table 1. Quantitative results on the private dataset (numbers in bold are the best indicator results).

Method	mIoU (%)	Mean F1 (%)	Mean precision (%)	Mean recall (%)
UNet	96.45	98.18	98.11	98.25
TransUNet	93.72	96.70	97.10	96.35
CGNet	95.57	97.71	97.54	97.87
SegFormer	95.11	97.46	97.31	97.61
GCNet	95.23	97.52	97.89	97.17
ContextNet	96.57	98.24	98.28	98.20
DABNet	96.32	98.11	97.86	98.36
FPENet	96.42	98.16	98.22	98.10
Ours	97.11	98.52	98.59	98.46