![](https://img-blog.csdnimg.cn/20201014180756754.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
Re-ID
cdy艳0917
这个作者很懒,什么都没留下…
展开
-
Beyond Part Models: Person Retrieval with Refined Part Pooling(PCB)
文章code年份2018-ECCV原创 2020-07-28 09:31:35 · 135 阅读 · 0 评论 -
Relation Network for Person Re-identification
文章code年份:AAAI 2020原创 2020-07-27 20:57:44 · 282 阅读 · 0 评论 -
Learning in the Frequency Domain-----2020CVPR
论文:Learning in the Frequency Domain https://arxiv.org/pdf/2002.12416.pdf年份:2020-CVPR 代码:https://github.com/calmevtime/DCTNet本文主要思想就是省略图像压缩/解压缩中计算量最大的步骤,直接利用频域特征来进行图像推理,减少系统中模块之间的数据传输量,从而提升系统性能。输入的数据量更小,深度神经网络在图像分类/分割任务上的精度反而提升了。该网络把RGB图像变换到DCT频率域,而不是直接下原创 2020-05-13 22:55:37 · 901 阅读 · 0 评论 -
Unsupervised Person Re-identification via Multi-label Classification
论文:Unsupervised Person Re-identification via Multi-label Classification年份:2020-CVPR文章目录1.简介2.导言3.相关工作无监督Person ReID(Unsupervised person ReID)无监督特征学习(Unsupervised feature learning)多标签分类4.方法4.1公式4.2 基...原创 2020-04-29 21:36:32 · 2667 阅读 · 6 评论 -
Self-Training With Progressive Augmentation for Unsupervised Cross-Domain Person Re-Identification
论文:Self-Training With Progressive Augmentation for Unsupervised Cross-Domain Person Re-Identificationhttps://arxiv.org/abs/1907.13315年份:2019ICCV代码:https://github.com/zhangxinyu-xyz/PAST-ReID文章目录简介...原创 2020-04-19 23:30:45 · 570 阅读 · 0 评论 -
Disentangled Image Matting
论文:Disentangled Image Mattinghttps://arxiv.org/abs/1909.04686年份:2019文章目录1.简介2.背景3.方法3.1Trimap Adaptation3.2Network Overview3.3 Multi-task Loss4.实验4.1Results on alphamatting.com4.2Results on Compositi...原创 2020-04-05 23:49:44 · 1191 阅读 · 0 评论 -
Vehicle Re-identification with Viewpoint-aware Metric Learning
论文:Vehicle Re-identification with Viewpoint-aware Metric Learninghttps://arxiv.org/abs/1910.04104时间:2019文章目录1.简述2.方法2.1网络架构2.2两个损失函数3实验1.简述本文提出一个车辆重识别网络 VANet,它能够学习具有视觉感知力的深度度量。VANet 将车辆重识别分为两个场景,即...原创 2020-04-04 23:15:31 · 457 阅读 · 1 评论 -
Foreground-aware Pyramid Reconstruction for Alignment-free Occluded Person Re-identification
论文:前景感知金字塔重建无对齐遮挡的行人重识别论文:https://arxiv.org/abs/1904.04975年份:2019ICCV文章目录1.背景2.方法FCNPyramid PoolingForeground Probability Generator3.之前研究方法的弊端本文提出了一种新颖的行人ReID遮挡鲁棒和无对准模型,并将其应用扩展到现实和拥挤的场景。所提出的模型首先...原创 2020-03-30 15:04:39 · 1399 阅读 · 0 评论 -
Peeking into occluded joints: A novel framework for crowd pose estimation
论文:Peeking into occluded joints: A novel framework for crowd pose estimation 代码:即将开源,后期更新年份:2020CVPR现有遮挡研究方法的内在问题缺陷一:它们根据视觉信息直接定位关节;然而,看不见的关节缺乏这一点。本论文解决方案:提出一个 Image-Guided Progressive GCN modul...原创 2020-03-31 09:07:52 · 980 阅读 · 0 评论 -
重识别论文讲解——双日一篇
决定读博了,所以对于重识别的论文打算好好研究一下,以后每偶数日更新一篇重识别论文的翻译与讲解,希望大家能够批评指正,我定及时修改,有代码的论文我也会简要看一下和大家分享,如果你发现我更新的论文有代码也请大家及时分享交流。...原创 2020-03-26 00:20:05 · 307 阅读 · 0 评论