1504.ICCVPartial Person Re-Identification 论文笔记
来源:互联网 发布:尊而光律所怎么样知乎 编辑:程序博客网 时间:2024/05/22 02:23
1504.ICCVPartial Person Re-Identification 论文笔记
作者针对之前一些模型都是假设给以完整行人检测图像probe情况下,进行行人重识别任务,独辟蹊径,主要解决reid最难处理的遮挡问题,研究给以不完整的行人图像块来进行ReID任务建模。
主要贡献是:
将应用在partial人脸识别中MSTR(multi-task sparse representation无需对齐,用稀疏特征表达,即词典的方法表示特征)方法迁移过来,在此基础上,添加约束减少错误对齐和匹配图像块的数目,引入图像块相似性评分机制,得到AMC(Ambiguity-sensitive Matching Classifier)模型,in order to model explicitly the diversity and ambiguity of the occlusion patterns in partial person re-id in our new sparse representation classification formulation. based on patch-level local-to-local
matching
另外,考虑了全局性的部件匹配,利用上空间布局信息,减少错匹配,提出SWM(sliding window matching )模型,based on a global-to-local matching by a detection-based matching model
然后,作者对上述模型的分类结果得分做加权融合,并做实验设置加权参数,验证了融合模型AWC-SWM模型AWC和SWC相较于不融合能取得更好的表现。
由于是提出的partial Re-ID模型,为与之前模型作对比,作者制作自己的数据集,在CAVIVR与i-LIDS扩展数据集上裁剪partial图像块,即得P-CAVIAR and P-iLIDS数据集,并将设计的AMC-SWM融合模型与MSTR/RDC/PRSVM/L1-norm/LFDA/KISSME/LADF/SRC做了不同数据集下的比较,验证了结果有效性。
另外,作者说明了带核的LFDA模型相较于LFDA模型对partialReID任务不会产生有益的贡献
最后作者提出改进方向:将解决多尺度图像匹配问题,以得分匹配形式引入到融合模型中
- 1504.ICCVPartial Person Re-Identification 论文笔记
- 论文笔记-Person Re-Identification Meets Image Search
- 论文笔记-Person Re-identification Past, Present and Future
- 论文笔记之---Person Re-identification in the Wild
- 论文笔记-Person Re-identification Past, Present and Future
- 论文笔记-Person Re-identification Past, Present and Future
- Person Re-Identification CVPR 2017 论文列表
- 论文导读(person re-identification)——Person Re-identification based on nonlinear ranking with difference
- 【论文笔记】In Defense of the Triplet Loss for Person Re-Identification
- 【论文笔记】Margin Sample Mining Loss: A Deep Learning Based Method for Person Re-identification
- 论文笔记Deeply-Learned Part-Aligned Representations for Person Re-Identification
- 1505.Deep Ranking for Person Re-identification via Joint Representation Learning论文笔记
- 论文阅读:Multi-Scale Triplet CNN for Person Re-Identification
- Person Re-identification Datasets
- Person Re-identification Datasets
- Person Re-identification Overview
- Person Re-identification Overview
- Person Re-identification Datasets
- 关于变量、对象(类)、String类的理解与区分
- 【51nod 1009】数字1的数量 【数位DP 模板】
- 记忆化搜索---最长子串
- 再见 2017,你好 2018
- JAVA JNI开发应用实例
- 1504.ICCVPartial Person Re-Identification 论文笔记
- Redis分布式锁的正确实现方式(Java版)
- 12月第一篇博客
- java多线程[7]:CyclicBarrier
- 深度学习之Caffe完全掌握:用C++开发(自定义)网络层
- TIPS
- 数据结构总结之KMP
- JavaScript 中的强制类型转换
- 【newcoder 】数圈圈 【数位DP 模板】