文章总览 - 191
DETECTIVE SAM:ADAPTIVE AI-IMAGE FORGERY LOCALIZATION
DETECTIVE SAM:ADAPTIVE AI-IMAGE FORGERY LOCALIZATION

21
MOODv2:Masked Image Modeling for Out-of-Distribution Detection
MOODv2:Masked Image Modeling for Out-of-Distribution Detection

发表于TIFS 2024,采用基于重构的方法实现 OOD 检测,发现无需微调,基于重构的预训练模型在OOD检测任务上表现优秀。

22
Segment Every Out-of-Distribution Object
Segment Every Out-of-Distribution Object

发表于CVPR 2024,探索了OOD检测的分割算法,本质上是将SAM模型加入到了OOD检测的分割任务,将传统OOD检测得到的异常分数图变成预测提示框,指导mask解码器定位OOD区域。

23
Rethinking Out-of-distribution (OOD) Detection:Masked Image Modeling is All You Need
Rethinking Out-of-distribution (OOD) Detection:Masked Image Modeling is All You Need

发表于CVPR 2023,采用基于重构的方法实现 OOD 检测,认为掩码图像建模任务可以顺利学习到图像自身的分布,实验结果达到了sota。

24
Towards Generalizable Detector for Generated Image
Towards Generalizable Detector for Generated Image

发表于NeurIPS 2025,完全基于自然图像训练的模型在生成图像检测,提出无需生成图片训练的方法DEnD,天才般的想法,与博主之前的某个想法不谋而合,无需训练便能达到sota,厉害!

25
ForgeLens:Data-Efficient Forgery Focus for Generalizable Forgery Image Detection
ForgeLens:Data-Efficient Forgery Focus for Generalizable Forgery Image Detection

发表于ICCV 2025。

26
RankMatch:Exploring the Better Consistency Regularization for Semi-supervised Semantic Segmentation
RankMatch:Exploring the Better Consistency Regularization for Semi-supervised Semantic Segmentation

发表于CVPR2024,其为每个像素设置代理模型以超越常规像素级一致性,还通过建模代理间关系实现等级感知的关联一致性,从而充分释放代理潜力。

27
Contrastive_learning_semantic_segmentation
Contrastive_learning_semantic_segmentation

28
CeCR:Cross-entropycontrastive replay for online class-incremental continual learning
CeCR:Cross-entropycontrastive replay for online class-incremental continual learning

29
Learning double-compression video fingerprints left from social-media platforms
Learning double-compression video fingerprints left from social-media platforms

30