Domain Adaptive Detection
2020-06-02 11:37:30 0 举报
域自适应目标检测研究现状
作者其他创作
大纲/内容
D&MKim et. al.CVPR
CDSemiSLYu et. al.arxiv
基于差异伪标签
DAPDGuo et. al.ICIP
DA-FasterChen et. al.CVPR
D2NCDArruda et. al.IJCNN
strong lcl-lvlweak glb-lvl
iFANZhuang et. al.AAAI
NLKhodabandeh et. al.ICCV
GAN+鲁棒伪标签一致性损失单阶段SSD
MMSCImg2ImgTLin et. al.ICIP
Sep.
TS-RPNCao et. al.INFFUS
先验知识来学习不变性残差特征恢复
Feb.
合成-真实glb-lvl baselcl-lvl istsbt-lvl msk
MAFHe et. al.ICCV
DASCRodriguez et. al.BMVC
rgn-miningrgn-lvl
混合性
img-lvlctg-awr ist-lvlctg-crlt ist-lvl
HTCNChen et. al.CVPR
GPAXu et.alCVPR
FAFRCNNWang et. al.CVPR
BfromADevaguptapu et. al.ICIP
多层对抗hier ftr-lvlprps ftr-lvl尺度减小模块
IR2VILiu et. al.CVPRW
可见-红外GAN
Oct.
伪标签校正
Mar.
GAN+伪标签迭代简单-复杂
不同于以往所有对抗工作多层对齐glb、istdmn-dmn翻译源域属性编码目标域
S2RDAZhang et. al.IJCNN
GAN+伪标签欠采样PL过采样TL迭代过大感受野降低生成质量
GAN+对抗两阶段适应加权训练ftr-lvl
前向后向循环梯度对齐熵正则进行域多样化
SWDASaito et. al.CVPR
图矩阵来表示proposal直接的关系
SCDAZhu et. al.CVPR
Apr.
MeanTeacherregion一致图内一致图间一致减小不同Aug的一致性损失学习不变性
RoyChowdhury et. al.CVPR
F&BCycleYang et. al.arxiv
风格迁移+伪标签
GAN插值校正迁移能力：lcl-rgnimgist
Aug.
基于重建 CycleGAN风格迁移
2018
伪标签精挑视频track
CurriSPLSoviany et. al.arxiv
DT-PLInoue et. al.CVPR
May.
Nov.
基于对抗域判别器
对抗训练判别域标签
红外-可见GAN+对抗
PDAHsu et. al.WACV
Prior-basedSindagi et. al.arxiv
伪标签+对抗单阶段
Day-NgtGAN
ST&ABRKim et. al.ICCV
June
MTORCai et. al.CVPR
SCLShen et. al.arxiv
P&FDAShan et. al.NeurComp
红外-可见GAN
堆叠辅助损失梯度分离策略-context reprs
GAN多样化+对抗
Dec.
CDNSu et. al.arxiv
可见-多谱自动迭代标注行人
2020
Few-shotimg-lvlist-lvl
2019
GAN+对抗pxl-lvlftr-lvl
July
Day-Ngt图像翻译新结构
0 条评论
回复 删除
下一页
