Muyun99's wiki Muyun99's wiki
首页
学术搬砖
学习笔记
生活杂谈
wiki搬运
资源收藏
关于
  • 分类
  • 标签
  • 归档
GitHub (opens new window)

Muyun99

努力成为一个善良的人
首页
学术搬砖
学习笔记
生活杂谈
wiki搬运
资源收藏
关于
  • 分类
  • 标签
  • 归档
GitHub (opens new window)
  • 论文摘抄

  • 论文阅读-图像分类

  • 论文阅读-语义分割

  • 论文阅读-知识蒸馏

  • 论文阅读-Transformer

  • 论文阅读-图卷积网络

  • 论文阅读-弱监督图像分割

    • Awesome weakly supervised semantic segmentation
    • Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation
    • Weakly-Supervised Image Semantic Segmentation Using Graph Convolutional Networks
    • Discriminative Region Suppression for Weakly-Supervised Semantic Segmentation
    • Weakly-Supervised Semantic Segmentation via Sub-category Exploration
      • Weakly-Supervised Semantic Segmentation via Sub-category Exploration
    • AffinityNet Learning Pixel level Semantic Affinity with Image level Supervision for Weakly Supervised Semantic Segmentation
    • Grad-CAM Visual Explanations from Deep Networks via Gradient-based Localization
    • Grad-CAM++ Improved Visual Explanations for Deep Convolutional Networks
    • Leveraging Auxiliary Tasks with Affinity Learning for Weakly Supervised Semantic Segmentation
    • Embedded Discriminative Attention Mechanism for Weakly Supervised Semantic Segmentation
    • Group-Wise Semantic Mining for Weakly Supervised Semantic Segmentation
    • Mining Cross-Image Semantics for Weakly Supervised Semantic Segmentation
    • NoPeopleAllowed The Three-Step Approach to Weakly Supervised SemanticSegmentation
    • Weakly Supervised Learning of Instance Segmentation with Inter-pixel Relations
    • Learning Deep Features for Discriminative Localization
    • Convolutional Random Walk Networks for Semantic Image Segmentation
    • Learning random-walk label propagation for weakly-supervised semantic segmentation
    • Puzzle-CAM Improved localization via matching partial and full features
    • Learning Visual Words for Weakly-Supervised Semantic Segmentation
    • 区域擦除 | Object Region Mining with Adversarial Erasing A Simple Classification to Semantic Segmentation Approach
    • CAM 扩散 | Tell Me Where to Look Guided Attention Inference Network
    • Self-Erasing Network for Integral Object Attention
    • Transformer CAM|Transformer Interpretability Beyond Attention Visualization
    • GETAM Gradient-weighted Element-wise Transformer Attention Map for Weakly-supervised Semantic segmentation
    • Class Re-Activation Maps for Weakly-Supervised Semantic Segmentation
  • 论文阅读-半监督图像分割

  • 论文阅读-带噪学习

  • 论文阅读-小样本学习

  • 论文阅读-自监督学习

  • 语义分割中的知识蒸馏

  • 学术文章搜集

  • 论文阅读-其他文章

  • 学术搬砖
  • 论文阅读-弱监督图像分割
Muyun99
2021-05-11

Weakly-Supervised Semantic Segmentation via Sub-category Exploration

# Weakly-Supervised Semantic Segmentation via Sub-category Exploration

# 单位:UC Merced, eBay, NEC Labs Ameriva, Google Research.

# 作者: Yu-Ting Chang, Qiaosong Wang, Wei-Chih Hung, Robinson Piramuthu, Yi-Hsuan Tsai, Ming-Hsuan Yang

# 发表:CVPR 2020

# 摘要

# 阅读

# 论文的目的及结论

# 论文的实验

image-20211029141541312

image-20211029141831905

image-20211029142157720

# 实验性能(3张表)
  • Tab.1 是和 AffinityNet 的比较
  • Tab.3 是和其他方法在每个类别上的性能比较
  • Tab.4 是和 SOTA 方法的性能比较

image-20211029141622685

image-20211029141629214

# 消融实验(2个表)

上面两张表是对子类数量 KKK 以及迭代次数的消融实验,最终K 取值为 10,round取值为3

image-20211029141548895

image-20211029141903659

image-20211029141844937

# 可视化结果(4张图)
  • Fig.3 展示了CAM的可视化结果
  • Fig.5 展示了聚类的结果
  • Fig.6 展示了特征空间上子类和父类的 t-SNE 结果,表明人这个子类别通常和其他的父类靠的很近,因为他们在同一张图里共同出现的概率很高
  • Fig.7 是量化结果

# 论文的方法

image-20211029141559455

# 论文的背景

# 总结

# 论文的贡献

将原始类别称为父类,在特征空间上对同父类样本的特征聚类后生成 KKK 个子类,聚类的结果作为子类的groundtruth,用于计算子类的loss。相当于做了一个额外的任务约束,来激励模型学到更多的语义信息。

# 论文的不足
# 论文如何讲故事

# 参考资料

  • https://arxiv.org/abs/2008.01183

  • https://github.com/Juliachang/SC-CAM

  • Refinement: We adopt the random walk method via affinity to refine the map as pixel-wise pseudo ground truths for semantic segmentation. Please refer to the repo of AffinityNet (opens new window) [1].

  • Segmentation network: We utilize the Deeplab-v2 framework [2] with the ResNet-101 architecture [3] as the backbone model to train the segmentation network. Please refer to the repo (opens new window).

上次更新: 2021/11/03, 23:35:28
Discriminative Region Suppression for Weakly-Supervised Semantic Segmentation
AffinityNet Learning Pixel level Semantic Affinity with Image level Supervision for Weakly Supervised Semantic Segmentation

← Discriminative Region Suppression for Weakly-Supervised Semantic Segmentation AffinityNet Learning Pixel level Semantic Affinity with Image level Supervision for Weakly Supervised Semantic Segmentation→

最近更新
01
Structured Knowledge Distillation for Semantic Segmentation
06-03
02
README 美化
05-20
03
常见 Tricks 代码片段
05-12
更多文章>
Theme by Vdoing | Copyright © 2021-2023 Muyun99 | MIT License
  • 跟随系统
  • 浅色模式
  • 深色模式
  • 阅读模式
×