Open-Vocabulary Semantic Segmentation on PASCAL VOC-20 (val)
98.5mIoUDPSeg
Evaluation Results
| Method | Links | |
|---|---|---|
| DPSegVenue=CVPR’25, Variant=Large2026.03 | 98.5 | |
| HyperCLIPVenue=CVPR’25, Variant=Large2026.03 | 98.3 | |
| ESCNetVenue=CVPR’25, Variant=Large2026.03 | 98.3 | |
| dinov3.segVariant=Large, Backbone=dinov3.txt2026.03 | 97.86 | |
| H-CLIPVenue=CVPR’25, Variant=Large2026.03 | 97.7 | |
| DEDOSVenue=ICCV’25, Variant=Large2026.03 | 97.6 | |
| SCANVenue=CVPR’24, Variant=Large2026.03 | 97.2 | |
| CAT-SegVenue=CVPR’24, Variant=Large2026.03 | 97 | |
| OVSNetVenue=ICCV’25, Variant=Large2026.03 | 96.9 | |
| MAFT+Venue=ECCV’24, Variant=Large2026.03 | 96.5 | |
| SEDVenue=CVPR’24, Variant=Large2026.03 | 96.1 | |
| FC-CLIPVenue=NeurIPS’23, Variant=Large2026.03 | 95.4 | |
| EOV-SegVenue=AAAI’25, Variant=Large2026.03 | 94.8 | |
| SANVenue=CVPR’23, Variant=Large2026.03 | 94.6 | |
| OVSegVenue=CVPR’23, Variant=Large2026.03 | 94.5 | |
| CLIPixtraining=false2026.07 | 87.8 | |
| ResCLIPtraining=false2026.07 | 87.1 | |
| CLIP-DINOisertraining=true2026.07 | 80.9 | |
| ClearCLIPtraining=false2026.07 | 80.9 | |
| SCLIPtraining=false2026.07 | 80.4 | |
| GroupViTtraining=true2026.07 | 79.7 | |
| TCLtraining=true2026.07 | 77.5 | |
| CLIPtraining=false2026.07 | 15.8 |