Tracking of Features and Edges
Joint Tracking of Features and Edges
1. LK光流
基本LK光流运动假设:
\]
一阶近似得到:
\]
由于Aperture problem,需要假设领域像素运动相同,并作为约束,便可以求解
\]
2. Horn-Schunck光流
\]
\(\lambda\)为正则项参数,相当于加了个平滑约束.
\(\nabla ^2u, \nabla ^2v\) 为\(u,v\)的拉普拉斯算子,可以近似为:
\]
领域\(u\)的均值来表示.
3. Joint Tracking
\]
E_S(i) = ((u_i-\hat{u}_i)^2+(v_i-\hat{v}_i)^2)
\]
\((\hat{u}_i,\hat{v}_i)^T\) 为期望的偏移量,可以通过任何一种方式获取.
Instead, we predict the motion displacement of a pixel by fitting an affine motion model
to the displacements of the surrounding features, which are inversely weighted according to their distance to the pixel.
We use a Gaussian weighting function on the distance, with σ = 10 pixels.
对于周围的特征拟合一个Affine变换来获取?
利用特征周围的特征点求解一个预测值:
- 直接利用领域内\((u,v)\)的平均值
特征选择:
\]
本文取: \(\eta=0.1\)
4. Unified Point-Edgelet Feature tracking
- 进一步优化,选取
Edgelet而不是边缘的点作为track的目标 - 预测的\((\hat{u},\hat{v})\)不是平均值,而是拟合一个Affine变换获得(u,v),并且拟合变换的权重根据距离和scale进行计算
5. \(u,v\)预测值如何计算
利用领域特征的\(u,v\)取加权来进行计算获得
6. 接下来工作
这些方法的思路都是利用点和边缘来互补操作,使得二者能够互相提升各自的缺陷,接下来基本参考joint_tracking的思路,但是不取平均值,而是进行加权操作,简单尝试.
7. 参考文献
- Birchfield S T , Pundlik S J . Joint tracking of features and edges
CVPR 2008 - Sundararajan K . Unified point-edgelet feature tracking[J]. Dissertations & Theses - Gradworks, 2011.
Tracking of Features and Edges的更多相关文章
- 深度学习Deep learning
In the last chapter we learned that deep neural networks are often much harder to train than shallow ...
- (转)A Beginner's Guide To Understanding Convolutional Neural Networks Part 2
Adit Deshpande CS Undergrad at UCLA ('19) Blog About A Beginner's Guide To Understanding Convolution ...
- Computer Vision Algorithm Implementations
Participate in Reproducible Research General Image Processing OpenCV (C/C++ code, BSD lic) Image man ...
- (zhuan) Notes on Representation Learning
this blog from: https://opendatascience.com/blog/notes-on-representation-learning-1/ Notes on Repr ...
- A successful Git branching model——经典篇
A successful Git branching model In this post I present the development model that I’ve introduced f ...
- 图像中的artifacts
artifacts 瑕疵 伪影(Artifacts) 伪影(Artifacts)-CT-基础术语 - 影像园 http://www.xctmr.com/baike/ct/c34b5413e305b45 ...
- 神奇的 ViewDragHelper,让你轻松定制拥有拖拽能力的 ViewGroup
为了吸引大家的注意力,先给大家看一张动图: 相信这种效果大家都见过吧?我第一次见到这样的效果时,心里也痒痒的,急于想实现这种功能,后来因为拖延症的问题,就一直没有去弄这件事.现在这段时间,工作比较轻闲 ...
- Convolution Fundamental I
Convolution Fundamental I Foundations of CNNs Learning to implement the foundational layers of CNN's ...
- [C6] Andrew Ng - Convolutional Neural Networks
About this Course This course will teach you how to build convolutional neural networks and apply it ...
随机推荐
- nginx 重发机制导致的重复扣款问题
问题: nginx 重发机制导制重复提交(客户还款,被扣俩笔款,前端调用一次,后端执行2次) proxy_next_upstream 语法: proxy_next_upstream error ...
- WordCount--实现字符,单词,代码统计
Github: https://github.com/whoNamedCody/WordCount PSP表格 PSP2.1 PSP阶段 预估耗时 (分钟) 实际耗时 (分钟) Planning 计 ...
- 快速排序Quick_Sort
快排——排序中的明星算法,也几乎是必须掌握的算法,这次我们来领略以下快排为何魅力如此之大. 快排主要有两种思路,分别是挖坑法和交换法,这里我们以挖坑法为例来进行介绍,交换法可以参考这篇博文.值得一提的 ...
- vue1 class style
- spring配置和映射文件
配置 <?xml version="1.0" encoding="UTF-8"?><beans xmlns="http://www. ...
- 17、生命周期-BeanPostProcessor在Spring底层的使用
17.生命周期-BeanPostProcessor在Spring底层的使用 bean赋值.注入其他组件.@Autowired注解.生命周期注解.@Async --都是 BeanPostProcesso ...
- HDU 6154 - CaoHaha's staff | 2017 中国大学生程序设计竞赛 - 网络选拔赛
/* HDU 6154 - CaoHaha's staff [ 构造,贪心 ] | 2017 中国大学生程序设计竞赛 - 网络选拔赛 题意: 整点图,每条线只能连每个方格的边或者对角线 问面积大于n的 ...
- Python 15__屏幕抓取
- 配置文件的属性ENC加密
转载:https://www.cnblogs.com/zqyx/p/9687136.html 在micro service体系中,有了config server,我们可以把配置存放在git.svn.数 ...
- springboot项目没错,但就是报红叉
1.报错原因: Description Resource Path Location TypeCannot change version of project facet Dynamic Web Mo ...