Table 4.

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

Method	Metrics	AP	AP50	AP75	APM	APL	AR	AR50	AR75	ARM	ARL
	OKS	0.883	0.925	0.897	0.899	0.897	0.935	0.927	0.916	0.890	0.943
VDN49	VDS	0.553	0.630	0.597	0.805	0.899	0.747	0.781	0.766	0.788	0.818
	OKS	0.880	0.918	0.899	0.886	0.887	0.924	0.952	0.942	0.911	0.930
Anam-Net61	VDS	0.419	0.476	0.465	0.736	0.990	0.703	0.769	0.750	0.738	0.991
	OKS	0.637	0.660	0.640	0.517	0.677	0.760	0.779	0.769	0.663	0.792
UNet62	VDS	0.232	0.292	0.243	0.957	0.706	0.540	0.596	0.558	0.658	0.709
	OKS	0.893	0.935	0.891	0.926	0.893	0.916	0.952	0.913	0.947	0.914
UNet++63	VDS	0.580	0.642	0.595	0.804	0.941	0.769	0.817	0.788	0.807	0.945
	OKS	0.013	0.031	0.007	0.024	0.014	0.159	0.279	0.125	0.179	0.158
SegNet64	VDS	0.024	0.038	0.023	0.152	0.258	0.134	0.173	0.135	0.149	0.255
	OKS	0.892	0.937	0.897	0.982	0.881	0.911	0.952	0.913	0.984	0.899
Enet65	VDS	0.569	0.692	0.586	0.789	0.941	0.745	0.846	0.750	0.791	0.945
	OKS	0.924	0.957	0.937	0.965	0.914	0.938	0.962	0.952	0.968	0.935
LEDNet66	VDS	0.605	0.718	0.623	0.828	0.946	0.761	0.846	0.769	0.833	0.950
	OKS	0.870	0.915	0.864	0.885	0.874	0.900	0.942	0.894	0.911	0.900
improved Unet++67	VDS	0.527	0.618	0.560	0.815	0.901	0.723	0.798	0.750	0.818	0.909
	OKS	0.940	0.965	0.942	0.935	0.947	0.951	0.971	0.952	0.942	0.963
improved U$_{}^2$-Net68	VDS	0.632	0.698	0.660	0.812	0.995	0.825	0.875	0.846	0.816	0.995
	OKS	0.893	0.929	0.939	0.926	0.949	0.957	0.982	0.960	0.937	0.971
AFFormer69	VDS	0.611	0.688	0.667	0.853	0.916	0.819	0.864	0.856	0.860	0.980
	OKS	0.923	0.958	0.925	0.933	0.937	0.956	0.968	0.932	0.943	0.981
EfficientSAM70	VDS	0.618	0.697	0.673	0.852	0.931	0.817	0.883	0.842	0.862	0.928
	OKS	0.941	0.970	0.942	0.932	0.947	0.961	0.981	0.962	0.942	0.971
PIN(Ours)	VDS	0.637	0.725	0.668	0.868	0.946	0.831	0.883	0.849	0.871	0.950

Significant values are in bold.