List your company
ResearchFormerly WizWand

Open-vocabulary Semantic Segmentation on PAS-20 Pascal VOC (test)

96.6mIoUCAT-Seg
15.16836.30957.4578.591Nov 27, 2023Apr 16, 2024Sep 4, 2024Jan 23, 2025Jun 13, 2025Nov 1, 2025Mar 22, 2026
Updated 7mo ago

Evaluation results

MethodLinks
CAT-SegVLM=ViT-L/14, Feature backbone=Swin-B, Training dataset=COCO-Stuff2023.1196.6
MAFT+semBackbone=ConvNeXt-L2026.0396.5
SEDVLM=ConvNeXt-L, Training dataset=COCO-Stuff2023.1196.1
MAFT+panBackbone=ConvNeXt-L2026.0395.6
SANVLM=ViT-L/14, Training dataset=COCO-Stuff2023.1195.5
OVRBackbone=ConvNeXt-L2026.0395.5
FC-CLIPVLM=ConvNeXt-L, Training dataset=COCO Panoptic [26]2023.1195.4
FC-CLIP*Backbone=ConvNeXt-L2026.0395.4
SANBackbone=ViT-L2026.0394.6
OVSegVLM=ViT-L/14, Feature backbone=Swin-B, Training dataset=COCO-Stuff+COCO Caption2023.1194.5
OVSegBackbone=ViT-L2026.0394.5
SEDVLM=ConvNeXt-B, Training dataset=COCO-Stuff2023.1194.4
OVRCOATBackbone=ConvNeXt-L2026.0394.1
SANVLM=ViT-B/16, Training dataset=COCO-Stuff2023.1194
CAT-SegVLM=ViT-B/16, Feature backbone=ResNet-101, Training dataset=COCO-Stuff2023.1193.7
OVSegVLM=ViT-B/16, Feature backbone=ResNet-101c, Training dataset=COCO-Stuff+COCO Caption2023.1192.6
DeOPVLM=ViT-B/16, Feature backbone=ResNet-101c, Training dataset=COCO-Stuff-1562023.1191.7
ZegFormerVLM=ViT-B/16, Feature backbone=ResNet-101, Training dataset=COCO-Stuff2023.1189.5
SimBaselineVLM=ViT-B/16, Feature backbone=ResNet-101, Training dataset=COCO-Stuff2023.1188.4
ZegFormerVLM=ViT-B/16, Feature backbone=ResNet-101, Training dataset=COCO-Stuff-1562023.1186.2
Han et al.VLM=ViT-B/16, Feature backbone=ResNet-101, Training dataset=COCO Panoptic [26]2023.1183.2
PACLVLM=ViT-B/16, Training dataset=GCC [40]+YFCC [44]2023.1172.3
OpenSegBackbone=ALIGN2026.0372.2
LSegVLM=ViT-B/32, Feature backbone=ViT-L/16, Training dataset=PASCAL VOC-152023.1152.3
GroupViTVLM=ViT-S/16, Training dataset=GCC [40]+YFCC [44]2023.1150.7
LSegVLM=ViT-B/32, Feature backbone=ResNet-101, Training dataset=PASCAL VOC-152023.1147.4
ZS3NetFeature backbone=ResNet-101, Training dataset=PASCAL VOC2023.1138.3
SPNetFeature backbone=ResNet-101, Training dataset=PASCAL VOC2023.1118.3