zzc_zhuyu的博客

ExchNet: A Unified Hashing Network for Large-Scale Fine-Grained Image RetrievalECCV 2020ExchNet：用于大规模细粒度图像检索的统一哈希网络文章目录ExchNet: A Unified Hashing Network for Large-Scale Fine-Grained Image Retrieval摘要1 引言2 相关研究2.1 细粒度图像检索2.2 深度哈希3 方法3.1 表示学习3.2 通过局部特征

Localization-Aware Adaptive Pairwise Margin Loss for Fine-Grained Image Recognition用于细粒度图像识别的定位感知自适应成对边距损失文章目录Localization-Aware Adaptive Pairwise Margin Loss for Fine-Grained Image Recognition摘要1 引言3 方法3.1 部分定位感知CutMix3.2 自适应成对边距损失正集合负集合4 实验4.1 SOTA4.

The Devil is in the Channels: Mutual-Channel Loss for Fine-Grained Image Classification（MC-Loss）2020,TIP文章目录The Devil is in the Channels: Mutual-Channel Loss for Fine-Grained Image Classification（MC-Loss）摘要1 引言3 MC-Loss3.1 判别性组件3.2 多样性组件4 实验SOTA比较消融摘要

Grad-CAM guided channel-spatial attention module for Fine-Grained Visual Classificationarxiv 2021文章目录Grad-CAM guided channel-spatial attention module for Fine-Grained Visual Classification摘要1 引言2 方法2.1 通道-空间注意力机制2.2 Grad-CAM2.3 指导损失3 实验摘要通道-空间注意力机制同时

Learning Attentive Pairwise Interaction for Fine-Grained Classification2020 AAAI。网络结构倒是不复杂，但是这么大的batch size要怎么跑起来。文章目录Learning Attentive Pairwise Interaction for Fine-Grained Classification摘要1 引言2 API-Net2.1 互矢量学习2.2 门向量2.3 成对交互2.4 训练与测试3 实验3.1 消融实验3.2

Your “Flamingo” is My “Bird”: Fine-Grained, or Not2021 CVPR，这还是第一篇看到还有从人工研究实验到模型的论文文章目录Your “Flamingo” is My “Bird”: Fine-Grained, or Not0 摘要1 引言2 相关研究3 人为研究数据参与者实验设置实验结果4 方法4.1 多粒度标签联合学习4.2 解耦和加强5 实验、结果和分析6 讨论7 结论0 摘要本文动机：在不同的专业水平下，如何为定制不同的细粒度定义。重新设

Learning Rich Part Hierarchies With Progressive Attention Networks for Fine-Grained Image Recognition2020 TIP，论文主要借鉴了MA-CNN。文章目录Learning Rich Part Hierarchies With Progressive Attention Networks for Fine-Grained Image Recognition1 引言3 方法3.1 多注意力模块及其损失3

Bi-Modal Progressive Mask Attention for Fine-Grained Recognition文章目录Bi-Modal Progressive Mask Attention for Fine-Grained Recognition摘要1 引言3 方法3.1 符号SAM模块QRM模块掩模模板3.2 视觉态PMA输入视觉表示PMA3.3 语言态PMA输入语言表示PMA3.4 特征聚合3.5 知识蒸馏4 实验2020 TIP摘要语言模态聚合被证明是一种改善视觉识别的技

Focus Longer to See Better: Recursively Refined Attention for Fine-Grained Image Classification文章目录Focus Longer to See Better: Recursively Refined Attention for Fine-Grained Image Classification摘要1 引言3 方法3.1 双分支架构全局分类局部分支3.2 分类损失3.3 联合特征4 实验4.2 分析4.3 消融实验

Re-rank Coarse Classification with Local Region Enhanced Features for Fine-Grained Image Recognition字节跳动AI实验室文章目录Re-rank Coarse Classification with Local Region Enhanced Features for Fine-Grained Image Recognition摘要1 引言3 方法3.1 全局特征的多层次损失3.2 弱监督区别性区域定位3

Interpretable and Accurate Fine-grained Recognition via Region Grouping通过区域分组实现可解释且准确的细粒度识别。根据，特征图获得每个语义部分的分配图，再计算出每个语义的特征向量，形成特征矩阵，后续变换和分类。文章目录Interpretable and Accurate Fine-grained Recognition via Region Grouping摘要1 引言2 相关研究3 方法3.1 部分分割和正则化部分分配发现部分正

Contrastively-reinforced Attention Convolutional Neural Network for Fine-grained Image Recognition个人感觉，这篇论文没有太多创新点，效果也一般般文章目录Contrastively-reinforced Attention Convolutional Neural Network for Fine-grained Image Recognition摘要1 引言2 相关研究3 方法3.1 ART模块3.2

Towards Robust Fine-grained Recognition by Maximal Separation of Discriminative Features通过 区分特征的最大分离 实现鲁棒的细粒度识别文章目录Towards Robust Fine-grained Recognition by Maximal Separation of Discriminative Features摘要1 引言2 相关研究3 预备阶段：解释对抗攻击4 方法4.1 架构4.2 区分特征分离5 实验6

Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patches文章目录Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patches0 摘要1 引言2 相关工作3 方法3.1 网络架构3.2 渐进式学习3.3 拼图生成4 实验4.1 细节4.2 比较SO

Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification from the Bottom Up文章目录Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification from the Bottom Up0 摘要1 引言2 相关工作3 方法3.1 总览3.2 弱监督目标检测和实例分割3.3 互补部分模

See Better Before Looking Closer: Weakly Supervised Data Augmentation Network for Fine-Grained Visual Classification文章目录See Better Before Looking Closer: Weakly Supervised Data Augmentation Network for Fine-Grained Visual Classification0 摘要1 引言2 相关工作3 方法3