Mean Shift 算法原理和在目标跟踪上的应用

Slides:



Advertisements
Similar presentations
663 Chapter 14 Integral Transform Method Integral transform 可以表示成如下的積分式的 transform  kernel Laplace transform is one of the integral transform 本章討論的 integral.
Advertisements

「方案設計」課程第六章 方案設計 — 目的與目標的設定 張 紉. 方案設計前三個要素 問題 界定 與 需求 評估 服務設計 資源 投入 運作 過程 服務 產出 「方案假設」
Dr. Baokun Li 经济实验教学中心 商务数据挖掘中心
第二章:随机变量 上节课内容 本节课内容 概率理论 随机变量及其分布 随机变量变换 常见分布族 多元随机向量的分布 概率公理及推论
教材:模式识别(第三版) 张学工编著 清华大学出版社
-Artificial Neural Network- Adaline & Madaline
Introduction To Mean Shift
An Adaptive Cross-Layer Multi-Path Routing Protocol for Urban VANET
Some Effective Techniques for Naive Bayes Text Classification
Applications of Digital Signal Processing
Platypus — Indoor Localization and Identification through Sensing Electric Potential Changes in Human Bodies.
Population proportion and sample proportion
多媒体搜索引擎 多媒体文档内容理解技术.
報告人:丁英智 資策會 網路多媒體研究所 11/3/2006
Image Retrieval Based on Fractal Signature
模式识别 Pattern Recognition
Manifold Learning Kai Yang
丁 承 國立交通大學經營管理研究所教授 成大統計68級 民國103年6月14日
非線性規劃 Nonlinear Programming
Seam Carving for Content-Aware Image Resizing
(Exec1) GIS 空间分析-使用ArcGIS (Exec1)
On Some Fuzzy Optimization Problems
Continuous Probability Distributions
第十章 基于立体视觉的深度估计.
Properties of Continuous probability distributions
Sampling Theory and Some Important Sampling Distributions
十一、簡單相關與簡單直線回歸分析(Simple Correlations and Simple Linear Regression )
Pattern Recognition Chapter1 Introduction.
光流法 (Optical Flow) 第八章 基于运动视觉的稠密估计 光流法 (Optical Flow)
Mean Shift Segmentation Algorithm
Step 1. Semi-supervised Given a region, where a primitive event happens Given the beginning and end time of each instance of the primitive event.
Depixelizing Pixel Art 像素风格画的矢量化
9.4 基于纹理的深度图重建.
机器人学基础 第四章 机器人动力学 Fundamentals of Robotics Ch.4 Manipulator Dynamics
Randomized Algorithms
Location Identification and Vehicle Tracking using VANET(VETRAC)
Interval Estimation區間估計
瞬态油漆混合器 练习 6.
Cyclic Hanoi问题 李凯旭.
塑膠材料的種類 塑膠在模具內的流動模式 流動性質的影響 溫度性質的影響
ICG 2018 Fall Homework1 Guidance
学习报告 —语音转换(voice conversion)
ZEEV ZEITIN Delft University of Technology, Netherlands
Source: IEEE Transactions on Image Processing, Vol. 25, pp ,
运动目标检测、提取 主 讲:刘 龙 2019年2月24日.
Chapter 5 Recursion.
資料結構 Data Structures Fall 2006, 95學年第一學期 Instructor : 陳宗正.
Version Control System Based DSNs
VIDEO COMPRESSION & MPEG
Monte Carlo模拟 引言(introduction) 均匀随机数的产生(Random number generation)
高性能计算与天文技术联合实验室 智能与计算学部 天津大学
相關統計觀念復習 Review II.
Summary for Chapters 24 摘要: 24章
3.5 Region Filling Region Filling is a process of “coloring in” a definite image area or region. 2019/4/19.
Chapter 04 流程能力與績效分析.
線性規劃模式 Linear Programming Models
Hashing Michael Tsai 2013/06/04.
五.連續變數及常態分佈 (Continuous Random Variables and Normal Distribution)
統計學回顧 區國強.
(四)标准差(standard deviation)
Hashing Michael Tsai 2017/4/25.
More About Auto-encoder
第六章 自組性類神經網路 類神經網路.
第四篇 螢幕擷取 CHAPTER 5 靜態螢幕擷取SnagIt
第三章 从概率分布函数的抽样 (Sampling from Probability Distribution Functions)
Monte Carlo模拟 引言(introduction) 均匀随机数的产生(Random number generation)
Chapter 0 Introduction to Medical Image Processing
之前都是分类的蒸馏很简单。然后从分类到分割也是一样,下一篇是检测的蒸馏
Gaussian Process Ruohua Shi Meeting
Presentation transcript:

Mean Shift 算法原理和在目标跟踪上的应用

Agenda Mean Shift Theory What is Mean Shift ? Density Estimation Methods Deriving the Mean Shift Mean shift properties Applications Clustering Discontinuity Preserving Smoothing Object Contour Detection Segmentation Object Tracking

Mean Shift Theory

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Mean Shift vector Objective : Find the densest region Distribution of identical billiard balls

Intuitive Description Region of interest Center of mass Objective : Find the densest region Distribution of identical billiard balls

研究现状 Mean shift算法是Fukunaga于1975年提出的,其含义即偏移的均值向量。随着Mean shift理论的发展,它的含义也发生了变化。现在一般是指一个迭代的步骤,即先算出当前点的偏移均值,移动该点到其偏移均值,然后以此为新的起始点,继续移动,直到满足一定的条件结束。Cheng Yizong定义了一族核函数 ,将Mean shift算法引入到计算机视觉领域。Bradski G R对Mean shift算法进行改进,发展建立了Camshift算法,将Mean shift方法扩展应用到了目标跟踪中来。

Mean shift的基本形式 给定d维空间 中的n个样本点,i=1,…,n,在点 的Mean Shift向量的基本形式定义为: 其中, 是一个半径为h的高维球区域, k表示在这n个样本点中,有k个点落入区域 中.

Mean shift的扩展 核函数: 代表一个d维的欧氏空间, 是该空间中的一个点,用一列向量表示。 的模 。 表示实数域。如果一个函数 存在一个剖面函数 ,即 剖面函数的性质: (1) 是非负的 ; (2) 是非增的; (3) 是分段连续的,并且

Kernel Density Estimation Various Kernels 在选定的空间中,x1…xn 是有限的样本点。 例: Epanechnikov Kernel Uniform Kernel (均匀核函数) Normal Kernel (高斯核函数)

核密度 估计 梯度 使用核函数 的形式: 得到 : 窗宽带宽

Computing The Mean Shift Yet another Kernel density estimation ! Simple Mean Shift procedure: Compute mean shift vector Translate the Kernel window by m(x)

Non-Rigid Object Tracking … …

Mean-Shift Object Tracking General Framework: Target Representation Choose a reference model in the current frame Choose a feature space Represent the model in the chosen feature space Current frame …

Mean-Shift Object Tracking General Framework: Target Localization Start from the position of the model in the current frame Search in the model’s neighborhood in next frame Find best candidate by maximizing a similarity func. Repeat the same process in the next pair of frames Current frame … Model Candidate

Mean-Shift Object Tracking Target Representation Choose a reference target model Represent the model by its PDF in the feature space Quantized Color Space Choose a feature space Kernel Based Object Tracking, by Comaniniu, Ramesh, Meer

Mean-Shift Object Tracking PDF Representation Target Model (centered at 0) Target Candidate (centered at y) Similarity Function:

Mean-Shift Object Tracking Finding the PDF of the target model model y candidate Target pixel locations A differentiable, isotropic, convex, monotonically decreasing kernel Peripheral pixels are affected by occlusion and background interference The color bin index (1..m) of pixel x Probability of feature u in model Probability of feature u in candidate Normalization factor Pixel weight Normalization factor Pixel weight

Mean-Shift Object Tracking Similarity Function Target model: Target candidate: Similarity function: 1 The Bhattacharyya Coefficient

Mean-Shift Object Tracking Target Localization Algorithm Start from the position of the model in the current frame Search in the model’s neighborhood in next frame Find best candidate by maximizing a similarity func.

Mean-Shift Object Tracking Approximating the Similarity Function Model location: Candidate location: Linear approx. (around y0) Independent of y Density estimate! (as a function of y)

Mean-Shift Object Tracking Maximizing the Similarity Function The mode of = sought maximum Important Assumption: One mode in the searched neighborhood The target representation provides sufficient discrimination

Mean-Shift Object Tracking Applying Mean-Shift The mode of = sought maximum Original Mean-Shift: Find mode of using Extended Mean-Shift: Find mode of using

Mean-Shift Object Tracking About Kernels and Profiles A special class of radially symmetric kernels: The profile of kernel K Extended Mean-Shift: Find mode of using

Mean-Shift Object Tracking Choosing the Kernel A special class of radially symmetric kernels: Epanechnikov kernel: Uniform kernel(单位均匀核函数): Extended Mean-Shift:

Mean-Shift Object Tracking Adaptive Scale Problem: The scale of the target changes in time The scale (h) of the kernel must be adapted Solution: Run localization 3 times with different h Choose h that achieves maximum similarity

完 谢谢