Multi-dimensional visual tracking using scatter search particle filter
Multi-dimensional visual tracking using scatter search particle filter
复制标题
DOI：
10.1016/j.patrec.2007.12.012
复制
发表时间：
2008-06
期刊：
影响因子：
--
通讯作者：
J. Pantrigo;Ángel Sánchez;A. S. Montemayor;A. Duarte
中科院分区：
文献类型：
--
作者：
J. Pantrigo;Ángel Sánchez;A. S. Montemayor;A. Duarte
多维视觉跟踪（MVT）问题包括视觉跟踪任务，其中系统状态由对应于多个模型组件和/或多个目标的大量变量定义。MVT问题可以建模为动态优化问题。在此背景下，我们提出了一种混合粒子滤波（PF）和散点搜索（SS）元启发式算法的算法，称为散点搜索粒子滤波（SSPF），其中散点搜索粒子滤波的优化策略嵌入到PF框架中。散点搜索是一种基于种群的元启发式算法，已成功地应用于若干复杂的组合优化问题。SS最具代表性的优化策略是方案组合和方案改进。组合阶段使解决方案能够共享有关问题的信息，从而产生更好的解决方案。改进阶段还可以通过探索给定解的邻域来获得更好的解。本文描述并评价了离散搜索粒子滤波器（SSPF）在MVT问题中的性能。具体来说，我们比较了几种最先进的基于pf的算法与SSPF算法在2D关节目标跟踪问题和2D多目标跟踪问题的不同实例中的性能。其中一些实例来自CVBase ' 06标准数据库。实验结果表明，该方法具有重要的性能增益和更好的跟踪精度。
Multi-dimensional visual tracking (MVT) problems include visual tracking tasks where the system state is defined by a high number of variables corresponding to multiple model components and/or multiple targets. A MVT problem can be modeled as a dynamic optimization problem. In this context, we propose an algorithm which hybridizes particle filters (PF) and the scatter search (SS) metaheuristic, called scatter search particle filter (SSPF), where the optimization strategies from SS are embedded into the PF framework. Scatter search is a population-based metaheuristic successfully applied to several complex combinatorial optimization problems. The most representative optimization strategies from SS are both solution combination and solution improvement. Combination stage enables the solutions to share information about the problem to produce better solutions. Improvement stage makes also possible to obtain better solutions by exploring the neighborhood of a given solution. In this paper, we have described and evaluated the performance of the scatter search particle filter (SSPF) in MVT problems. Specifically, we have compared the performance of several state-of-the-art PF-based algorithms with SSPF algorithm in different instances of 2D articulated object tracking problem and 2D multiple object tracking. Some of these instances are from the CVBase’06 standard database. Experimental results show an important performance gain and better tracking accuracy in favour of our approach.