Table 3.

Performance comparison of different models for semantic segmentation (Significant values are highlighted in bold).

Model	Precision	Recall	F1 Score	IoU	Accuracy	Inference Speed
UNet	0.7942	0.8053	0.8177	0.6887	0.9327	166 FPS
SegNet	0.8269	0.7720	0.7866	0.6802	0.9345	145 FPS
PSPNet	0.6631	0.7257	0.6756	0.5553	0.8777	95 FPS
DeepLabv3	0.8162	0.8537	0.8329	0.6900	0.9481	89 FPS
Swin-UNet	0.7829	0.8040	0.7925	0.6827	0.9511	98.92 FPS
Lightweight ViT-based Model	0.6029	0.5823	0.5848	0.4708	0.8428	150 FPS
Proposed PSPEdgeWeedNet	0.9399	0.9335	0.9358	0.8797	0.9335	86 FPS