List your company
ResearchFormerly WizWand

Semantic Segmentation on COCO 2014 (val)

61.8mIoUFSSS
43.49648.2485357.752Mar 6, 2019May 10, 2020Jul 16, 2021Sep 21, 2022Nov 26, 2023Jan 31, 2025Apr 8, 2026
Updated 6mo ago

Evaluation results

MethodLinks
FSSSBackbone=ResNet-101, Supervision=F2023.0461.8
ModuSegTrain status=Training-Free, Method category=Multi-stage2026.0456.7
SemPLeS2024.0156.1
FMA-WSSS2024.0155.4
G-RAM-SAM2024.0154.6
AttentionShots=5, Multi-scale (MS)=true2019.0351.6
WeakTrBackbone=EVA-02-S, Sup.=I, large-scale datasets=true2023.0451.1
SSRTrain status=Required, Method category=Single-stage2026.0450.6
UniMatch* + S4MCBackbone=Xception-65, Partition protocol=1/32 (3697)2023.0850.58
WeakTrBackbone=ViT-S, Sup.=I, large-scale datasets=true2023.0450.3
ExCELTrain status=Required, Method category=Single-stage2026.0450.3
CANetShots=1, Multi-scale (MS)=true2019.0349.9
UniMatch*Backbone=Xception-65, Partition protocol=1/32 (3697)2023.0849.8
S2CTrain status=Required, Method category=Multi-stage2026.0449.8
VPLTrain status=Required, Method category=Multi-stage2026.0449.8
AttentionShots=5, Multi-scale (MS)=false2019.0349.7
MARSBackbone=ResNet-101, Supervision=I2023.0449.4
Mask-AvgShots=5, Multi-scale (MS)=false2019.0349.2
Feature-AvgShots=5, Multi-scale (MS)=false2019.0348.9
WeakTrBackbone=DINOv2-S, Sup.=I, large-scale datasets=true2023.0448.9
UniMatch* + S4MCBackbone=Xception-65, Partition protocol=1/64 (1849)2023.0847.98
POTBackbone=ResNet101, Sup.=I + L, large-scale datasets=false2023.0447.9
POTTrain status=Required, Method category=Multi-stage2026.0447.9
UniMatch*Backbone=Xception-65, Partition protocol=1/64 (1849)2023.0847.8
OTPLTrain status=Required, Method category=Multi-stage2026.0447.6
MoReTrain status=Required, Method category=Single-stage2026.0447.4
WeakCLIPTrain status=Required, Method category=Multi-stage2026.0447.4
PCREBackbone=ViT-B, Sup.=I, large-scale datasets=true2023.0447.2
PCRETrain status=Required, Method category=Single-stage2026.0447.2
WeCLIPBackbone=ViT, Supervision=I+L, Stage=single-stage weakly supervised, CRF=true2024.0647.1
WeCLIPBackbone=CLIP-ViT-B, Sup.=I + L, large-scale datasets=true2023.0447.1
WeCLIPTrain status=Required, Method category=Single-stage2026.0447.1
WeakTrBackbone=DeiT-S, Sup.=I, large-scale datasets=false2023.0446.9
CM-GLasso2026.0446.82
SFCBackbone=ResNet101, Supervision=Image-level supervision only2024.0146.8
SFCBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0446.8
FFRBackbone=ViT-B, Sup.=I, large-scale datasets=true2023.0446.8
CPALTrain status=Required, Method category=Multi-stage2026.0446.8
SeCoType=single-staged, Backbone=ViT-B/16, Supervision=image labels2024.0246.7
MuP-VSSBackbone=ResNet38, Sup.=I, large-scale datasets=false2023.0446.6
MulP-VSS2026.0446.6
RS+EPMBackbone=ResNet-101, Supervision=I2023.0446.4
RSEPMBackbone=R101, Supervision=I, Framework=Multi-stage2022.0446.4
WeCLIPBackbone=ViT, Supervision=I+L, Stage=single-stage weakly supervised, CRF=false2024.0646.4
CANetShots=1, Multi-scale (MS)=false2019.0346.3
Mask-ORShots=5, Multi-scale (MS)=false2019.0346.2
PC2SegBackbone=Xception-65, Partition protocol=1/32 (3697)2023.0846.1
KTSEBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0445.9
RSBackbone=R101, Supervision=I, Framework=Multi-stage2022.0445.8
Text2MaskBackbone=R101, Supervision=I+L (Image and Language)2023.0945.7
PCSSBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0445.7
FBR (AMN-based)Sup.=I, Backbone=ResNet1012024.0645.6
CDLBackbone=ResNet101, Supervision=I, Stage=multi-stage weakly supervised2024.0645.5
DiGBackbone=ResNet38, Sup.=I, large-scale datasets=false2023.0445.5
CLIP-ESBackbone=ResNet-101, Segmentation Network=DeepLabV2, Supervision Type=Image-level + Language supervision (I+L)2022.1245.4
CLIP-ESType=multi-staged, Backbone=ResNet101, Supervision=image labels + external data2024.0245.4
CLIP-ESBackbone=ViT+Res101, Supervision=I+L, Stage=multi-stage weakly supervised2024.0645.4
CLIP-ESBackbone=R101, Supervision=I+L (Image and Language)2023.0945.4
CPF-CTEBackbone=ViT-B/162026.0145.4
CLIP-ESBackbone=ResNet101, Sup.=I + L, large-scale datasets=false2023.0445.4
CTIBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0445.4
CLIP-ES2026.0445.4
CLIP-ESTrain status=Required, Method category=Multi-stage2026.0445.4
ACRBackbone=WR38, Supervision=I, Framework=Multi-stage2022.0445.3
ACRBackbone=ResNet38, Supervision=I, Stage=multi-stage weakly supervised2024.0645.3
ACRSup.=I, Backbone=ResNet382024.0645.3
ACRBackbone=ResNet38, Sup.=I, large-scale datasets=false2023.0445.3
MCTformer+Train status=Required, Method category=Multi-stage2026.0445.2
BECOBackbone=ViT, Supervision=I, Stage=multi-stage weakly supervised2024.0645.1
BECOSup.=I, Backbone=ResNet1012024.0645.1
BECOBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0445.1
BECO2026.0445.1
RECAMSup.=I, Backbone=ResNet1012024.0645
ViT-PCMSup.=I, Backbone=ViT-B/162024.0645
ViT-PCMBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0445
AEFTSup.=I, Backbone=ResNet382024.0644.8
Yoon et al.Backbone=ResNet38, Sup.=I, large-scale datasets=false2023.0444.8
AMNBackbone=ResNet101, Supervision=I., Pre-training=ImageNet2022.0344.7
AMNBackbone=ResNet-101, Segmentation Network=DeepLabV2, Supervision Type=Image-level (I)2022.1244.7
AMNBackbone=ResNet-101, Supervision=I2023.0444.7
SANCEBackbone=ResNet-50, Supervision=I2023.0444.7
AMNBackbone=R101, Supervision=I, Framework=Multi-stage2022.0444.7
SANCEBackbone=R101, Supervision=I, Framework=Multi-stage2022.0444.7
AMNBackbone=ResNet101, Supervision=Image-level supervision only2024.0144.7
AMNSup.=I, Backbone=ResNet1012024.0644.7
AMNBackbone=R101, Supervision=I (Image-level)2023.0944.7
AMNBackbone=ResNet101, Sup.=I, large-scale datasets=false2023.0444.7
DUPL2024.0144.6
DuPLBackbone=ViT-B, Sup.=I, large-scale datasets=true2023.0444.6
DuPLTrain status=Required, Method category=Single-stage2026.0444.6
SASBackbone=ViT-B/162026.0144.5
WeakTr2024.0144.4
DIALBackbone=ViT-B, Sup.=I + L, large-scale datasets=true2023.0444.4
USAGEBackbone=ResNet101, Supervision=I, Stage=multi-stage weakly supervised2024.0644.3
L2G*Seg. Architecture=V2, Supervision=I.+S., Backbone=ResNet-1012022.0444.2
L2GBackbone=ResNet-101, Segmentation Network=DeepLabV2, Supervision Type=Image-level + Saliency maps (I+S)2022.1244.2
L2GBackbone=ResNet-101, Supervision=I+S2023.0444.2
L2GBackbone=R101, Supervision=I+S, Framework=Multi-stage2022.0444.2
L2GType=multi-staged, Backbone=ResNet101, Supervision=image labels + saliency maps2024.0244.2
L2GBackbone=ResNet101, Supervision=I+S, Stage=multi-stage weakly supervised2024.0644.2