Table 7.

Comparison with the different approaches in terms of the evaluation of OKS and VDS metrics on MPT dataset.

Method	Metrics	AP	AP50	AP75	APM	APL	AR	AR50	AR75	ARM	ARL
	OKS	0.796	0.800	0.800	0.796	0.796	0.799	0.805	0.810	0.810	0.799
VDN49	VDS	0.717	0.744	0.732	0.805	0.799	0.737	0.756	0.748	0.811	0.806
	OKS	0.798	0.809	0.798	0.798	0.798	0.803	0.818	0.801	0.799	0.803
Anam-Net61	VDS	0.699	0.733	0.712	0.813	0.811	0.737	0.760	0.748	0.817	0.817
	OKS	0.817	0.820	0.820	0.817	0.817	0.823	0.826	0.826	0.821	0.823
UNet62	VDS	0.693	0.730	0.707	0.796	0.849	0.743	0.768	0.76	0.801	0.853
	OKS	0.799	0.801	0.801	0.799	0.799	0.803	0.805	0.805	0.800	0.803
UNet++63	VDS	0.724	0.755	0.733	0.790	0.820	0.740	0.764	0.748	0.796	0.826
	OKS	0.041	0.089	0.036	0.045	0.045	0.195	0.309	0.202	0.200	0.195
SegNet64	VDS	0.042	0.060	0.048	0.064	0.224	0.127	0.160	0.136	0.059	0.221
	OKS	0.797	0.800	0.800	0.797	0.797	0.799	0.801	0.801	0.801	0.799
Enet65	VDS	0.679	0.703	0.703	0.778	0.807	0.712	0.727	0.727	0.782	0.814
	OKS	0.825	0.830	0.830	0.825	0.825	0.826	0.830	0.830	0.830	0.826
LEDNet66	VDS	0.713	0.756	0.726	0.791	0.843	0.733	0.768	0.744	0.797	0.846
	OKS	0.780	0.781	0.781	0.780	0.780	0.788	0.789	0.789	0.788	0.788
improved Unet++67	VDS	0.733	0.754	0.742	0.788	0.830	0.748	0.760	0.752	0.795	0.835
	OKS	0.799	0.800	0.800	0.799	0.799	0.808	0.809	0.809	0.800	0.808
improved U$_{}^2$-Net68	VDS	0.714	0.745	0.730	0.811	0.798	0.734	0.752	0.744	0.815	0.805
	OKS	0.852	0.862	0.865	0.863	0.856	0.873	0.869	0.853	0.872	0.847
AFFormer69	VDS	0.736	0.813	0.764	0.889	0.876	0.793	0.829	0.801	0.885	0.896
	OKS	0.867	0.874	0.861	0.857	0.866	0.863	0.859	0.867	0.890	0.863
EfficientSAM70	VDS	0.737	0.812	0.765	0.869	0.876	0.787	0.832	0.798	0.881	0.907
	OKS	0.875	0.881	0.871	0.879	0.875	0.877	0.881	0.877	0.891	0.877
PIN(Ours)	VDS	0.795	0.834	0.795	0.888	0.901	0.816	0.848	0.815	0.895	0.906

Significant values are in bold.