Object tracking based on two-dimensional PCA

In this paper, we present a novel object tracking method based on two-dimensional PCA. The low quality of images and the changes of the object appearance are very challenging for the object tracking. The representation of the training features is usually used to solve these challenges. Two-dimensional PCA (2DPCA) based on the image covariance matrix is constructed directly using the original image matrices. An appearance model is presented and its likelihood estimation has been established based on 2DPCA representation in this paper. Compared with the state-of-the-art methods, our method has higher reliability and real-time property. The performances of the proposed tracking method are quantitatively and qualitatively shown in experiments.     

An object tracking method based on guided filter for night fusion image

Online object tracking is a challenging problem as it entails learning an effective model to account for appearance change caused by intrinsic and extrinsic factors. In this paper, we propose a novel online object tracking with guided image filter for accurate and robust night fusion image tracking. Firstly, frame difference is applied to produce the coarse target, which helps to generate observation models. Under the restriction of these models and local source image, guided filter generates sufficient and accurate foreground target. Then accurate boundaries of the target can be extracted from detection results. Finally timely updating for observation models help to avoid tracking shift. Both qualitative and quantitative evaluations on challenging image sequences demonstrate that the proposed tracking algorithm performs favorably against several state-of-art methods.     

利用黑洞原理改进稀疏外观模型跟踪算法

为了解决跟踪漂移问题,提出了一种利用黑洞原理改进的稀疏外观模型目标跟踪算法,用来提高目标跟踪的鲁棒性。利用黑洞原理从目标模板中搜索聚类中心来降低目标模板数量。通过学习分类器用于构造目标特征;用黑洞原理获取模板字典表示目标;采用高斯分布运动模型获取目标样本,在贝叶斯框架下根据观测模型获取最优目标位置实现跟踪。不同视频序列被用于改进的稀疏外观模型跟踪算法和其他先进目标跟踪算法进行仿真实验。实验结果表明,实现了目标跟踪的目的,有效地降低了目标局部遮挡问题的影响,提高了目标跟踪精度。     

基于简单线性迭代聚类超像素的meanshift跟踪

为了增强目标跟踪算法在被跟踪目标发生运动位移、遮挡、形变、相似物体干扰等情况下的鲁棒性,提出利用超像素构建目标外观模型,将外观模型与候选区域进行匹配,获取候选区域当中目标超像素,并用Meanshift算法确定目标中心点的跟踪算法。仿真实验选取Benchmark库当中在运动位移、遮挡、形变、相似物体干扰方面具有代表性的视频Girl和FaceOcc1。该算法在视频Girl中的跟踪成功率和跟踪精度为0.601、0.856,比对比实验的经典算法当中跟踪效果最好的KCF算法的成功率和精度分别高0.059和0.084;在视频FaceOcc1中跟踪成功率和精度仅次于KCF。表明该跟踪算法在受到相似物体干扰和目标遮挡时具有良好的鲁棒性。     

多目标跟踪中目标数量准确识别算法

由于场景中存在目标遮挡与重叠的影响,以及检测时产生的分割错误,同一个目标往往被识别为若干运动区域,或者多个目标被识别为同一个运动区域,从而导致目标数量识别错误。为了解决这一问题,提出了一种结合融合与分离操作的理论来分析视频的方法,将多目标跟踪的问题转化成为一个找寻后验极值的问题。采用图论的方法表示观测结果,利用图像序列中目标运动轨迹和外观相似度的信息,将多目标跟踪的问题归结为找寻图中多个最优路径的问题。该方法采用了滑动窗口框架以便统计固定数目帧的信息。实验结果表明该方法能够应对实际中发生的上述现象,达到了准确识别多目标数量的目的。     

基于支持向量机增量学习和LPBoost的人体目标再识别算法

摄像机间目标关联是无重叠视域多摄像机目标持续跟踪的关键.提出了一种只利用人体目标外观,完全不依赖于空时关系的人体目标再识别算法,利用识别结果直接进行跨摄像机间人体目标关联,而不依赖于目标的捕获时间和路径限制.对跟踪视频前景图像序列提取互补性视觉单词树直方图和全局颜色直方图二种特征,采用支持向量机增量学习在线训练二种特征的人体外观辨别模型,再利用多类线性规划增强算法对二种特征的支持向量机模型进行在线自适应融合.实验结果表明,本文算法具有较强的在线学习能力,能增量式表达人体目标辨别性外观模型,特征融合后的模型区别性更强,有效地降低多方面条件变化的影响,获得了高识别率,且能够实现快速实时实现,相对于现有方法有了明显提升.     

Robust visual tracking based on multi-cue integration

Traditional color-based mean shift tracking algorithm is unable to accurately track the object. To address this problem, we present an improved tracking algorithm. The improved tracker integrates the color and motion cues which characterize the appearance and motion information of the object, respectively. These two visual cues can complement each other and make for more precise target localization. Experiments show that the proposed tracking algorithm has better performance than the traditional mean shift tracker.     

Mean shift tracker with Chaotic Artificial Bee Colony and Space Variant Resolution

A novel mean shift algorithm is proposed for object tracking in this paper. The mean shift procedure with Chaotic Artificial Bee Colony (Chaotic ABC) and Space Variant Resolution (SVR) of human visual system is utilized for adaptation of the target acceleration and estimation of the target's scale and orientation changes. In order to test the effectiveness and robustness of our proposed method, two groups of experiments were carried out and the related results of the proposed mean shift tracker with Chaotic ABC and SVR (MS-Chaotic ABC&SVR) are compared with three other algorithms, which demonstrate that our proposed approach is most robust and effective in solving object tracking problems than the others.     

Visual tracking based on fragment-based correntropy induced metric

This paper proposes a fast and effective template-based visual tracking method based on a novel distance metric. First, a fragment-based correntropy induced metric (FCIM) is proposed, which exploit the strengths from both the correntropy method and the fragment scheme to handle observation noise (especially, gross errors caused by occlusion). Second, the proposed FCIM method is integrated into the Bayesian inference framework for solving the tracking problem. In addition, a simple on-line template update scheme is introduced to capture the appearance change of the object during the tracking processing. Experimental results and discussions demonstrate that the proposed method is better than other popular algorithms.     

基于自适应背景的目标跟踪算法研究

随着现在的社会发展以及经济进步,我国的科学技术方面发展迅速,特别是在技术监控方面更是突飞猛进。为了更好的对目标遮挡影响进行降低,我国在这方面主要依据自适应的技术发展背景下提出目标跟踪计算法,用来完善我国的监督控制技术。这种计算方式第一是根据对观察目标的基本外观形态进行的鉴定与跟踪,将其自身的运动量进行平均计算;其次是根据时空的运行方向与特征进行跟踪目标的计算,建立比较完善整体的运行模型,再根据这个运动模型以及整体的状态对监督目标进行检测与控制,这期间就会形成一种遮挡掩膜。对于掩膜是一种将程序数据等绘制成光刻板,在程序使用期间非常可靠,并且制造成本比较低,使用方便;最后是在不同的使用情况下将不同参数进行收集,自动的适应运动模型的运行。针对这种计算方式的实验主要是利用两种在国际上经常使用的CAVIAR、York数据进行测试,并且根据这两种数据对测试的精准度与多重目标跟踪等进行评定,检测跟踪的整体性能。通过多方面的研究表明这种方式的跟踪的性能非常好,并且还能很好的将跟踪目标的鲁棒性进行遮挡。     

基于直方图比的背景加权的Mean Shift目标跟踪算法

针对Mean Shift跟踪算法目标模型中背景像素所造成的目标跟踪定位的偏差，提出了一种基于直方图比的背景加权的目标表示方法。该方法使用由目标核直方图和背景直方图的对数似然比值推导出的隶属度因子作为权值，通过只对目标模型进行加权变换，而不变换目标候选模型的方法，增强了目标和背景的可分性，有效减少了跟踪过程中背景像素的干扰，提高了目标定位的准确性。 仿真实验结果表明算法的有效性。     

相位一致性图像及其在目标跟踪中的应用(英文)

针对传统实时相关跟踪方法对照度变化敏感的问题,提出了一种基于相位一致性图像的相关跟踪方法.利用相位一致性函数值在[0,1]区间内且无量纲、对图像的亮度和对比度具有不变性等特点,首先对原始图像进行相位一致性检测,得到相位一致性图像,再利用MAD(Minimum Absolute Difference)等相关跟踪算法在相位一致性图像中对目标进行跟踪运算.对可见光和红外图像的实验表明,在图像的亮度和对比度发生剧烈变化的情况下,算法仍能保持对目标的稳定跟踪.该方法可用于解决传统实时相关跟踪方法普遍存在的因照度变化导致跟踪点漂移甚至跟踪失败的问题.     

Joint tracking algorithm using particle filter and mean shift with target model updating

Roughly, visual tracking algorithms can be divided into two main classes: deterministic tracking and stochastic tracking. Mean shift and particle filter are their typical representatives, respectively. Recently,a hybrid tracker, seamlessly integrating the respective advantages of mean shift and particle filter (MSPF)has achieved impressive success in robust tracking. The pivot of MSPF is to sample fewer particles using particle filter and then those particles are shifted to their respective local maximum of target searching space by mean shift. MSPF not only can greatly reduce the number of particles that particle filter required, but can remedy the deficiency of mean shift. Unfortunately, due to its inherent principle, MSPF is restricted to those applications with little changes of the target model. To make MSPF more flexible and robust, an adaptive target model is extended to MSPF in this paper. Experimental results show that MSPF with target model updating can robustly track the target through the whole sequences regardless of the change of target model.     

基于自适应粒子滤波的红外目标跟踪

为有效解决非线性环境中的红外目标跟踪问题,提出一种自适应粒子滤波目标跟踪算法.建立了目标加权概率模型.在滤波过程中,提出双过程粒子重抽样方法,形成对抽样粒子集的自适应调节,有效地解决了粒子退化问题.用实际红外图像序列做了实验.结果表明,在非线性环境下用该方法得到的红外目标跟踪结果优于用传统粒子滤波和扩展卡尔曼滤波算法获得的结果.     

真实场景下视频运动目标自动提取方法

视频运动目标跟踪逐渐成为研究热点并应用到军事民用等领域,为了能够从真实场景中快速准确地提取视频跟踪单运动目标或多运动目标,提出了一种新的运动目标自动提取方法。首先通过自适应阈值获得滤波后的相邻帧差值图像。其次,为了消除差值图像中噪声的影响,标记此二值图像的连通像素来检测出运动目标所在的区域,并与边缘检测出的空间信息结果比较得到运动目标模型。最后,将图像分成若干区域,在每个分区域内依次连接每个运动目标模型的最外围轮廓点,由此构成目标闭合轮廓。利用得到的连续边界,对运动目标进行提取。实验结果表明,该算法能够有效地自动提取速度不同的单运动目标,同时能够提取多运动目标。     

Robust visual tracking with contiguous occlusion constraint

Visual tracking plays a fundamental role in video surveillance, robot vision and many other computer vision applications. In this paper, a robust visual tracking method that is motivated by the regularized \(\ell\)1 tracker is proposed. We focus on investigating the case that the object target is occluded. Generally, occlusion can be treated as some kind of contiguous outlier with the target object as background. However, the penalty function of the \(\ell\)1 tracker is not robust for relatively dense error distributed in the contiguous regions. Thus, we exploit a nonconvex penalty function and MRFs for outlier modeling, which is more probable to detect the contiguous occluded regions and recover the target appearance. For long-term tracking, a particle filter framework along with a dynamic model update mechanism is developed. Both qualitative and quantitative evaluations demonstrate a robust and precise performance.     

Eigenspace-based tracking for feature points

Feature point tracking deals with image streams that change over time. Most existing feature point tracking algorithms only consider two adjacent frames at a time, and forget the feature information of previous frames. In this paper, we present a new eigenspace-based tracking method that learns an eigenspace representation of training features online, and finds the target feature point with Gauss-Newton style search method. A coarse-to-fine processing strategy is introduced to handle large affine transformations. Several simulations and experiments on real images indicate the effectiveness of the proposed feature tracking algorithm under the conditions of large pose changes and temporary occlusions.     

基于柯西分布的视频图像序列背景建模和运动目标检测

提出了一种用于视觉监视系统的基于柯西分布的发光模型的光照不变变化检测方法.假定视频图像序列中每个背景图像像素点灰度观测值的时序变化由白噪声引起,利用建立的初始化背景高斯统计模型对每帧图像进行归一化,得到了背景像灰度比值的分布符合标准柯西分布的结论,解决了柯西分布的模型参量估计问题.在变化检测的基础上,YCbCr颜色空间的亮度、色调和饱和度被用来识别和消除由阴影和反光等引起的变化区域.结果表明,提出的背景建模方法对场景中各种光线变化、小的背景扰动等噪声具有稳健性,可以较为可靠地检测前景目标,识别和去除阴影和反光.     

一种结合粒子滤波和张量子空间的目标跟踪算法

由于传统的子空间方法易于丢失图像目标的二维特性,为此本文提出了一种新颖的自适应目标跟踪算法,通过张量的方式建立目标的外观模型——张量子空间,利用在线学习的方法更新其外观模型,同时,利用目标仿射运动的先验信息,通过粒子滤波自适应地跟踪运动目标,并将获得的最优目标观测作为新数据反馈回子空间更新.此外,为了保证子空间更新能获得精确且紧致的目标子空间表达,引入动态部分函数滤除样本野点.实验结果表明,本文提出的自适应目标跟踪方法具有较强的鲁棒性,对于存在姿态变化、短时遮挡和光照变化等情况下均可有效地跟踪目标.     

A novel infrared small moving target detection method based on tracking interest points under complicated background

Infrared moving target detection is an important part of infrared technology. We introduce a novel infrared small moving target detection method based on tracking interest points under complicated background. Firstly, Difference of Gaussians (DOG) filters are used to detect a group of interest points (including the moving targets). Secondly, a sort of small targets tracking method inspired by Human Visual System (HVS) is used to track these interest points for several frames, and then the correlations between interest points in the first frame and the last frame are obtained. Last, a new clustering method named as R-means is proposed to divide these interest points into two groups according to the correlations, one is target points and another is background points. In experimental results, the target-to-clutter ratio (TCR) and the receiver operating characteristics (ROC) curves are computed experimentally to compare the performances of the proposed method and other five sophisticated methods. From the results, the proposed method shows a better discrimination of targets and clutters and has a lower false alarm rate than the existing moving target detection methods.