List your company
ResearchFormerly WizWand

Object Detection on PASCAL VOC + COCO-20 2007 (test)

83.2APJoint Training
43.78454.01764.2574.483Jun 1, 2021Jul 27, 2021Sep 22, 2021Nov 17, 2021Jan 13, 2022Mar 10, 2022May 6, 2022
Updated 8mo ago

Evaluation results

MethodLinks
Joint TrainingBackbone=Sparse R-CNN, Exemplars=false2022.0583.2——52.9
ROSETTABackbone=Sparse R-CNN, Exemplars=false2022.0582.3——49.5
Joint TrainingBackbone=Faster R-CNN, Exemplars=false2022.0581.6——48.8
Fine-tuningBackbone=Sparse R-CNN, Exemplars=false2022.0581.2——35.7
ROSETTABackbone=Faster R-CNN, Exemplars=false2022.0580.5——48.6
Fine-tuningBackbone=Faster R-CNN, Exemplars=false2022.0579.5——23.2
AFDBackbone=Faster R-CNN, Exemplars=false2022.0577.1——27.8
AFDBackbone=Faster R-CNN, Exemplars=true2022.0575.2——36.8
EWCBackbone=Faster R-CNN, Exemplars=false2022.0575——27.2
MASBackbone=Faster R-CNN, Exemplars=false2022.0574.8——28.1
EWCBackbone=Faster R-CNN, Exemplars=true2022.0573.4——32.2
MASBackbone=Faster R-CNN, Exemplars=true2022.0573.4——32.7
LwF DetectionBackbone=Faster R-CNN, Exemplars=true2022.0573——26.6
Attention DistillationBackbone=Faster R-CNN, Exemplars=true2022.0573——28.5
Feature DistillationBackbone=Faster R-CNN, Exemplars=true2022.0572.4——26.9
Dynamic Thresholding and Re-weightingbackbone=ResNet-50, multi_scale_training=true2021.0656.179.661.2—
Humble Teacherbackbone=ResNet-50, multi_scale_training=true2021.0654.481.3——
Dynamic Thresholding and Re-weightingbackbone=ResNet-502021.065477.659.1—
Unbiased Teacherbackbone=ResNet-50, multi_scale_training=true2021.0650.378.8——
Instant-Teachingbackbone=ResNet-502021.0649.77954.1—
ISMTbackbone=ResNet-502021.0649.677.7——
STACbackbone=ResNet-502021.064679.1——
Supervisedbackbone=ResNet-502021.0645.376.347.5—
CSDbackbone=ResNet-502021.06—75.1——