Learning a Discriminative Null Space for Person Re-identification CVPR 2016

运用KNFST进行降维，用于行人再识别。

原理部分  NFST KNFST

实验部分

数据库

VIPeR: 论文提供了提取特征后的数据。

VIPeR contains 632 identities and each has two images captured outdoor from two views with distinct view angles. All images are scaled to 128 × 48 pixels。

632 people’s images are randomly divided into two equal halves, one for training and the other for testing. This is repeated for 10 times and the averaged performance is reported.  A和B两个场景中各选择一半作为训练集合，另外一半作为测试集合。

A，B的训练集合集成在一起形成训练集。A的测试作为gallary, B的测试作为probe.  最终通过训练得到对特征空间的投影矩阵（隐式的，核方法），计算样本在低维空间中的维度，然后利用欧式距离进行识别，评估采用累积匹配曲线， CMC Cumulated Matching Characteristics。

时间： 2024-10-10 20:08:58

Learning a Discriminative Null Space for Person Re-identification CVPR 2016的相关文章

Kernel Null Space Methods for Novelty Detection CVPR 2013

Kernel Null Space Methods for Novelty Detection  CVPR 2013 B是St零空间的正交补空间的基向量,可以通过对Xt进行标准正交化或者对St进行PCA得到. 经证明,B'SwB的零空间的维数为C-1. 中心化的核矩阵 这里需要用到又核矩阵计算总体散度矩阵的非零特征向量. 先计算中心化的核矩阵的特征向量,总体散度矩阵的特征向量可以用(13)导出. v相当于是减去均值后的样本的线性组合系数.(特征向量由列样本的线性组合而成) Xw=Fi*(I-L)

Jordan 11 Space Jam to come back in 2016

Fans who’ve been following Air Jordan’s earlier releases already know by what the company is after because of its Jordan releases 2016. One of these simple pieces may be the Jordan 11 Low Bredsscheduled to become launched in June.his Jordans 11 model

(zhuan) Deep Reinforcement Learning Papers

Deep Reinforcement Learning Papers A list of recent papers regarding deep reinforcement learning. The papers are organized based on manually-defined bookmarks. They are sorted by time to see the recent papers first. Any suggestions and pull requests

目标跟踪算法综述

转自  https://www.zhihu.com/question/26493945 作者:YaqiLYU 第一部分:目标跟踪速览 先跟几个SOTA的tracker混个脸熟,大概了解一下目标跟踪这个方向都有些什么.一切要从2013年的那个数据库说起..如果你问别人近几年有什么比较niubility的跟踪算法,大部分人都会扔给你吴毅老师的论文,OTB50和OTB100(OTB50这里指OTB-2013,OTB100这里指OTB-2015,50和100分别代表视频数量,方便记忆): Wu Y, L

A Brief Review of Supervised Learning

There are a number of algorithms that are typically used for system identification, adaptive control, adaptive signal processing, and machine learning. These algorithms all have particular similarities and differences. However, they all need to proce

Machine Learning Algorithms Study Notes(4)—无监督学习（unsupervised learning）

1    Unsupervised Learning 1.1    k-means clustering algorithm 1.1.1    算法思想 1.1.2    k-means的不足之处 1.1.3    如何选择K值 1.1.4    Spark MLlib 实现 k-means 算法 1.2    Mixture of Gaussians and the EM algorithm 1.3    The EM Algorithm 1.4    Principal Components

transfer learning

作者:刘诗昆链接:https://www.zhihu.com/question/41979241/answer/123545914来源:知乎著作权归作者所有.商业转载请联系作者获得授权,非商业转载请注明出处. Transfer learning 顾名思义就是就是把已学训练好的模型参数迁移到新的模型来帮助新模型训练数据集.就跟其他知友回答的那样,考虑到大部分数据或任务是存在相关性的,所以通过transfer learning我们可以将已经学到的parameter 分享给新模型从而加快并优化模型的学

Applied Deep Learning Resources

Applied Deep Learning Resources A collection of research articles, blog posts, slides and code snippets about deep learning in applied settings. Including trained models and simple methods that can be used out of the box. Mainly focusing on Convoluti

机器学习基石第三讲 Types of Learning

一.Learning with Different Output Space Y 二.Learning with Different Data Label yn 监督式学习,每个训练样本都有对应的label,相当于每个x都有一个y对应 非监督式学习,每个训练样本是没有label的,需要电脑自己划分,类似于聚类,只有x而没有y 有些时候样本过于庞大,没法完全标签,只能对其中一部分标签,所以诞生了半监督式学习 增强式学习,很难明确的定义一个label,很难给x一个明确的y,比如21点,很难给要牌或者