List your company
ResearchFormerly WizWand

Semantic Segmentation on Pascal Context 459 classes (val)

21.6mIoUMAFT+
4.2328.74113.2517.759Apr 3, 2023Jul 14, 2023Oct 24, 2023Feb 3, 2024May 15, 2024Aug 25, 2024Dec 5, 2024
Updated 8mo ago

Evaluation results

MethodLinks
MAFT+VLM Backbone=ConvNeXt-L, Ensemble Operation=false2024.0821.6
FC-CLIPtraining dataset=COCO Panoptic2023.0818.2
FC-CLIPVLM Backbone=ConvNeXt-L, Ensemble Operation=true2024.0818.2
SANtraining dataset=COCO Stuff [5]2023.0817.1
MAFTVLM Backbone=ConvNeXt-L, Ensemble Operation=false2024.0817
SAM w/ Mask-AdapterBackbone=SAM-H, Setting=Open-vocabulary2024.1216.8
SCANVL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1216.7
MAFTVLM Backbone=ViT-L, Ensemble Operation=false2024.0816.2
SANVL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1215.7
SANVLM Backbone=ViT-L, Ensemble Operation=false2024.0815.7
ODISEtraining dataset=COCO Panoptic2023.0814.5
ODISEVL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1214.5
ODISEVLM Backbone=ViT-L, Ensemble Operation=false2024.0814.5
ODISEtraining dataset=COCO Panoptic + COCO Caption, variant=caption2023.0813.8
SCANVL-Model=CLIP VIT-B/16, Training Dataset=COCO2023.1213.2
MAFTVL-Model=CLIP ViT-B/16, Training Dataset=COCO2023.1212.8
SAM w/ MAFTBackbone=SAM-H, Setting=Open-vocabulary2024.1212.8
FC-CLIPVLM Backbone=ConvNeXt-L, Ensemble Operation=false2024.0812.7
SANVL-Model=CLIP VIT-B/16, Training Dataset=COCO2023.1212.6
OVSegtraining dataset=COCO Stuff [5]2023.0812.4
OVSegVL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1212.4
OVSegVLM Backbone=ViT-L, Ensemble Operation=false2024.0812.4
OpenSegVL-Model=ALIGN EN-B7, Training Dataset=COCO+Loc. Narr.2023.1212.2
OpenSegVLM Backbone=ALIGN, Ensemble Operation=false2024.0812.2
OVSegVL-Model=CLIP VIT-B/16, Training Dataset=COCO2023.1211
SimSeg†VL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1210.2
MaskCLIPtraining dataset=COCO Panoptic2023.0810
MaskCLIPVL-Model=CLIP VIT-L/14, Training Dataset=COCO2023.1210
SimSeg†VL-Model=CLIP VIT-B/16, Training Dataset=COCO2023.129.7
DeOPBackbone=R101c, Training dataset=COCO-Stuff-156, Decoupled network=true, Number of Passes=12023.049.4
OpenSegtraining dataset=COCO Panoptic + COCO Caption2023.089
OpenSegVL-Model=ALIGN EN-B7, Training Dataset=COCO2023.129
SimSegVL-Model=CLIP VIT-B/16, Training Dataset=COCO2023.128.7
OpenSegBackbone=R101, Training dataset=COCO + Loc. Narr., Decoupled network=false, Number of Passes=12023.047.9
LSeg+training dataset=COCO Stuff [5]2023.087.8
LSeg+VL-Model=ALIGN EN-B7, Training Dataset=COCO2023.127.8
SimBaseBackbone=R101c, Training dataset=COCO-Stuff-156, Decoupled network=true, Number of Passes=N'2023.047
OpenSegBackbone=R101, Training dataset=COCO, Decoupled network=false, Number of Passes=12023.046.5
OpenSegVL-Model=ALIGN RN101, Training Dataset=COCO2023.126.5
SAM w/ CLIPBackbone=SAM-H, Setting=Open-vocabulary2024.126.4
LSeg+VL-Model=ALIGN RN101, Training Dataset=COCO2023.125.2
GroupViTtraining dataset=GCC [75]+YFCC [79]2023.084.9