Hand-object segmentation on EgoHOS out-of-domain (test) benchmark leaderboard · SOTA2 Research

Hand-object segmentation on EgoHOS out-of-domain (test)

0.9721Left Hand AccuracyInterFormer

Evaluation Results

Method	Links
InterFormer2026.02		0.9721	0.971	0.7922	0.6819	0.5874	0.8009
CaRe-Ego2026.02		0.9709	0.9669	0.723	0.609	0.4628	0.7465
Mask2former2026.02		0.9705	0.9638	0.6439	0.6418	0.3978	0.7236
ANNEXE2026.02		0.9703	0.968	0.7657	0.6035	0.5235	0.7662
UperNet2026.02		0.9689	0.96	0.6483	0.5459	0.2783	0.6803
Maskformer2026.02		0.9558	0.961	0.7053	0.6049	0.4652	0.7384
Multi-UNet2026.02		0.9475	0.958	0.4516	0.4789	0.1675	0.6007
SCTNet2026.02		0.9462	0.9092	0.4914	0.2747	0.1712	0.5585
Seq2026.02		0.8783	0.8598	0.5717	0.4385	0.5476	0.6592
Segformer2026.02		0.868	0.7944	0.2177	0.0956	0.1167	0.4185