Language-guided Grasp Detection and Segmentation on OCID VLG (Multi-Split) benchmark leaderboard · SOTA2 Research

Language-guided Grasp Detection and Segmentation on OCID VLG (Multi-Split)

0.854J@1LGGD

Evaluation Results

Method	Links
LGGD2025.12		0.854	0.902	0.831	0.98	0.937	0.913	0.667	0.261
CLIP + MHSA2025.12		0.825	0.877	0.713	0.932	0.844	0.545	0.277	0.001
EfficientGrasp + CLIP2025.12		0.792	0.824	0.802	0.874	0.718	0.431	0.321	0.156
CROG2025.12		0.772	—	0.811	0.969	0.948	0.872	0.641	0.164
SSG + CLIP2025.12		0.335	—	0.336	0.356	0.356	0.355	0.355	0.328
GT-Grounding2025.12		0.287	—	—	—	—	—	—	—
Det-Seg + CLIP2025.12		0.281	—	0.29	0.272	0.209	0.175	0.172	0.16
GT-Masks + CLIP2025.12		0.119	—	0.35	0.35	0.35	0.35	0.35	0.35
GLIP + SAM2025.12		0.107	—	0.303	0.347	0.341	0.335	0.286	0.117
GR-ConvNet + CLIP2025.12		0.097	0.154	0.313	0.21	0.116	0.055	0.024	0.005
SAM + CLIP2025.12		0.072	—	0.257	0.293	0.285	0.274	0.227	0.091