Table 2.

Comparative analysis with classic semantic segmentation models. Bold font indicates the best values.

Model	Precision(%)	Recall(%)	Accuracy(%)	MIoU(%)	F1-score(%)
SegFormer	72.32	41.18	99.10	67.33	52.48
DeepLabV3+	68.11	66.12	98.99	63.57	67.10
BiSeNetV2	60.84	58.39	98.86	57.06	59.59
Fast-SCNN	49.90	58.55	98.78	56.76	53.88
DANet	70.60	62.68	98.96	61.00	66.40
VDP_UNet	71.44	65.63	99.27	75.63	68.41