List your company
ResearchFormerly WizWand

Semantic Segmentation on COCO-St. (test)

45.4mIoUCAA
11.49620.29829.137.902Jan 19, 2021Sep 1, 2021Apr 14, 2022Nov 25, 2022Jul 8, 2023Feb 18, 2024Sep 30, 2024
Updated 8mo ago

Evaluation results

MethodLinks
CAABackbone=EfficientNet-B7, Multi-scale prediction=true2021.0145.4
HRNetV2 + OCR + RMIBackbone=HRNetV2, Multi-scale prediction=true2021.0145.2
DRANMulti-scale prediction=true2021.0141.2
HRNetV2 + OCRBackbone=HRNetV2, Multi-scale prediction=true2021.0140.5
SegCLIPTraining Dataset=COCO [60], CC12M [56], Backbone=ViT-B/16, Additional Labels=Captions, VFM=CLIP2024.0926.5
ZeroSegTraining Dataset=ImageNet-1K [61], Backbone=ViT-B/16, Additional Labels=-, VFM=CLIP, SAM2024.0923.6
ZeroSegTraining Dataset=ImageNet-1K [61], Backbone=ViT-B/16, Additional Labels=-, VFM=CLIP, DINO2024.0922.2
PixelCLIPTraining Dataset=5% SA-1B [13] (0.5M), Backbone=ConvNeXt-B, Additional Labels=-, VFM=CLIP, SAM2024.0921.4
ZeroSegTraining Dataset=ImageNet-1K [61], Backbone=ViT-B/16, Additional Labels=-, VFM=CLIP2024.0920.2
PixelCLIPTraining Dataset=5% SA-1B [13] (0.5M), Backbone=ConvNeXt-B, Additional Labels=-, VFM=CLIP, DINO2024.0920.2
TCLTraining Dataset=CC3M, CC12M [56], Backbone=ViT-B/16, Additional Labels=Captions, VFM=CLIP2024.0919.6
CLIPTraining Dataset=WIT-400M [15], Backbone=ViT-B/16, Additional Labels=-, VFM=-2024.0916.5
GroupViTTraining Dataset=CC12M [56], YFCC15M [57], Backbone=ViT-S/16, Additional Labels=-, VFM=-2024.0915.3
OpenCLIPTraining Dataset=LAION-2B [59], Backbone=ConvNeXt-B, Additional Labels=-, VFM=-2024.0912.8