![](https://img-blog.csdnimg.cn/20201014180756927.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
文献
文章平均质量分 94
一只黑猩猩
这个作者很懒,什么都没留下…
展开
-
Semi-supervised Learning for COVID-19 Image Classification via ResNet
Semi-supervised Learning for COVID-19 Image Classification via ResNet27 Feb 2021Lucy Nwuso, Xiangfang Li, Lijun Qian, Senior Member, IEEE, Seungchan Kim, Member, IEEE, XishuangDong, Member, IEEE文章目录Semi-supervised Learning for COVID-19 Image Classifica原创 2021-03-28 21:11:39 · 311 阅读 · 0 评论 -
Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers
文章目录一、研究背景二、研究重点三、研究方法3.1 图像序列化3.2 Encoder3.3 Decoder3.3.1 Naive upsampling (Naive)3.3.2 Progressive UPsampling (PUP)3.3.3 Multi-Level feature Aggregation (MLA)四、实验及结果4.1 实验相关参数4.2 ADE20K结果4.3 Pascal Context结果4.4 Cityscapes结果五、消融实验六、其他可视化6.1 注意力机制6.2 位置向量6原创 2021-01-11 21:29:22 · 2979 阅读 · 9 评论 -
Transformer讲解以及在CV领域的应用
文章目录Transformer一、思想和框图二、实现细节2.1 Encoder2.2 Decoder2.3 Self-Attention2.4 Multi-Headed Attention2.5 Positional Encoding2.6 Layer normalization三、应用任务和结果3.1 NLP领域3.2 CV领域3.2.1 检测DETR3.2.2 分类ViT3.2.3 分割SETR3.2.4 Deformable-DETR四、优点及分析五、缺点及分析六、参考文献Transformer原创 2021-01-10 14:47:24 · 29648 阅读 · 9 评论 -
Self2Self With Dropout: Learning Self-Supervised Denoising From Single Image【使用单张图像进行自监督学习去噪】
文章目录Self2Self With Dropout: Learning Self-Supervised Denoising From Single Image【使用单张图像进行自监督学习去噪】一、相关概念1.1 监督学习(Supervised learning)1.2 无监督学习(Unsupervised learning)1.3 自监督学习(Self-Supervised learning)二、研究概述2.1 研究背景2.2 研究现状与问题2.3 研究内容2.4 研究结论三、研究方法3.1 模型3.2原创 2020-10-15 19:54:22 · 4790 阅读 · 5 评论 -
Simple Black-box Adversarial Attacks【简易的黑盒对抗攻击】
Simple Black-box Adversarial Attacks【简易的黑盒对抗攻击】Chuan Guo, Jacob R. Gardner, Yurong You, Andrew Gordon Wilson, Kilian Q. WeinbergerGitHub:https://github.com/cg563/simple-blackbox-attack一、相关概念1.1 对抗攻击(Adversarial Attack)基于深度学习的图像分类网络,大多是在精心制作的数据集下进行训原创 2020-07-31 08:41:23 · 6913 阅读 · 3 评论