List your company
ResearchFormerly WizWand

Object Detection on MS-COCO 5% labeled protocol

32.1mAPDual Pseudo-label Polishing Framework
17.924821.604925.28528.9651May 10, 2020Sep 20, 2020Jan 31, 2021Jun 13, 2021Oct 24, 2021Mar 6, 2022Jul 17, 2022
Updated 8mo ago

Evaluation results

MethodLinks
Dual Pseudo-label Polishing Framework2022.0732.1
Virtual Category Learning (with tricks)Configuration=Training tricks in Soft Teacher2022.0732.05
Soft Teacher2022.0730.74
Soft TeacherRemark=ICCV 20212022.0730.74
Virtual Category LearningConfiguration=Same settings as Unbiased Teacher2022.0730.31
Region Uncertainty QuantificationStrong data augmentation=true2021.1128.96
Combating Noise2022.0728.96
Unbiased Teacher2022.0728.27
Unbiased TeacherRemark=ICLR 20212022.0728.27
Humble TeacherStrong data augmentation=true2021.1127.7
Humble Teacher2022.0727.7
Humble TeacherRemark=CVPR 20212022.0727.7
Instant-Teaching*Backbone=R50-FPN, co-rectify scheme=true2021.0326.75
Instant-TeachingStrong data augmentation=true2021.1126.75
Instant Teaching2022.0726.75
Instant-Teaching*Remark=CVPR 20212022.0726.75
Interactive2022.0726.37
Unbiased TeacherStrong data augmentation=true2021.1126.3
Instant-TeachingBackbone=R50-FPN, co-rectify scheme=false2021.0325.5
Instant-TeachingRemark=CVPR 20212022.0725.5
STACBackbone=ResNet-50 with FPN, Augmentation=C+{B,G}+Cutout, Training Data=Semi-supervised2020.0524.38
STACBackbone=R50-FPN2021.0324.38
STACStrong data augmentation=true2021.1124.38
STAC2022.0724.38
STACRemark=arxiv 20202022.0724.38
Supervised†Backbone=ResNet-50 with FPN, Augmentation=C+{B,G}+Cutout, Training Data=Labeled data only2020.0521.18
Supervised2022.0720.92
CSD†Backbone=R50-FPN2021.0318.9
CSDStrong data augmentation=false2021.1118.63
CSD2022.0718.63
SupervisedBackbone=ResNet-50 with FPN, Augmentation=Standard, Training Data=Labeled data only2020.0518.47
SupervisedBackbone=R50-FPN2021.0318.47
Supervised2022.0718.47