Lightweight Real-Time Image Semantic Segmentation Network Based on Multi-Resolution Hybrid Attention Mechanism

Table 6

Comparison with advanced results on the Camvid dataset. Input size is resolution.

Model	Resolution	GFLOPs	Parameters	FPS	mIoU
ERFNet [16]	360480	2.07 M	8.43	133	65.0
DABNet [7]	360480	0.76 M	—	104	66.4
EDANet [34]	360480	0.68 M	8.97	—	66.4
ASFNet [8]	360480	5.38 M	5.07	220	68.0
FRFNet [32]	360480	4.02 M	—	225	68.2
MHANet(ours)	360480	5.42 M	4.76	257	70.07