List your company
ResearchFormerly WizWand

Semantic Segmentation on PASCAL VOC with background category VOC21 2012

86.8mIoUCoCo-SAM3
15.66434.13252.671.068Dec 2, 2024Feb 25, 2025May 21, 2025Aug 14, 2025Nov 7, 2025Jan 31, 2026Apr 27, 2026
Updated 6mo ago

Evaluation results

MethodLinks
CoCo-SAM3Paradigm=Training-free SAM32026.0486.8
ReMEParadigm=Training-free CLIP-VFM2026.0482.2
SAM3Paradigm=Training-free SAM32026.0481.9
CorrCLIPParadigm=Training-free CLIP-VFM2026.0476.7
CorrCLIPBackbone=ViT-H, Training=Training-free, Instance Guidance=SAM22026.0475.5
DouCBackbone=ViT-H, Training=Training-free, Instance Guidance=SAM22026.0474.3
DouCBackbone=ViT-L, Training=Training-free, Instance Guidance=SAM22026.0473.1
DouCBackbone=ViT-B, Training=Training-free, Instance Guidance=SAM22026.0472.2
CorrCLIPBackbone=ViT-B, Training=Training-free, Instance Guidance=SAM22026.0472
CorrCLIPBackbone=ViT-L, Training=Training-free, Instance Guidance=SAM22026.0471.2
TridentBackbone=ViT-H, Training=Training-free2026.0470.8
CLIPerBackbone=ViT-L, Training=Training-free2026.0469.8
FSABackbone=ViT-H, Training=Training-free2026.0467.9
RF-CLIPParadigm=Training-free CLIP-Only2026.0467.2
TridentParadigm=Training-free CLIP-VFM2026.0467.1
TridentBackbone=ViT-B, Training=Training-free2026.0467.1
Talk2DINOParadigm=Training-based2026.0465.8
CASSParadigm=Training-free CLIP-VFM2026.0465.8
CASSBackbone=ViT-B, Training=Training-free2026.0465.8
SC-CLIPBackbone=ViT-L, Training=Training-free2026.0465
ProxyCLIPBackbone=ViT-H, Training=Training-free2026.0465
DIH-CLIPBackbone=ViT-B, Training=Training-free2026.0464.2
SFPParadigm=Training-free CLIP-Only2026.0463.9
SFPBackbone=ViT-B, Training=Training-free2026.0463.9
FSABackbone=ViT-B, Training=Training-free2026.0463.7
CLIP-DINOiserParadigm=Training-based2026.0462.1
FSABackbone=ViT-L, Training=Training-free2026.0461.8
ProxyCLIPBackbone=ViT-L, Training=Training-free2026.0460.6
SCLIPData Size=400M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1259.1
SCLIPParadigm=Training-free CLIP-Only2026.0459.1
SCLIPBackbone=ViT-B, Training=Training-free2026.0459.1
NACLIPParadigm=Training-free CLIP-Only2026.0458.9
NACLIPBackbone=ViT-B, Training=Training-free2026.0458.9
CDAMBackbone=ViT-B, Training=Training-free2026.0458.7
ProxyCLIPParadigm=Training-free CLIP-VFM2026.0458.6
FreeDABackbone=ViT-L, Training=Training-free2026.0455.4
ResCLIPBackbone=ViT-L, Training=Training-free2026.0454.1
COSMOS w/ SCLIPData Size=30M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1252.8
ClearCLIPParadigm=Training-free CLIP-Only2026.0451.8
FreeDAParadigm=Training-free CLIP-VFM2026.0451.8
TCLParadigm=Training-based2026.0451.2
CLIPtraseParadigm=Training-free CLIP-Only2026.0450.9
GroupViTParadigm=Training-based2026.0450.4
CaRParadigm=Training-free CLIP-Only2026.0448.6
MaskCLIPData Size=400M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1243.1
COSMOSData Size=30M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1229.6
OpenCLIPData Size=1B, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1222.5
OpenCLIPData Size=2B, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1219.8
CLIPData Size=400M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1218.8
CLIPParadigm=Training-free CLIP-Only2026.0418.6
OpenCLIPData Size=400M, Zero-shot=true, Vision Encoder Architecture=ViT-B/162024.1218.4