Mean shift spectral clustering

In recent years there has been a growing interest in clustering methods stemming from the spectral decomposition of the data affinity matrix, which are shown to present good results on a wide variety of situations. However, a complete theoretical understanding of these methods in terms of data distributions is not yet well understood. In this paper, we propose a spectral clustering based mode merging method for mean shift as a theoretically well-founded approach that enables a probabilistic interpretation of affinity based clustering through kernel density estimation. This connection also allows principled kernel optimization and enables the use of anisotropic variable-size kernels to match local data structures. We demonstrate the proposed algorithm's performance on image segmentation applications and compare its clustering results with the well-known Mean Shift and Normalized Cut algorithms.     

Mean shift is a bound optimization

We build on the current understanding of mean shift as an optimization procedure. We demonstrate that, in the case of piecewise constant kernels, mean shift is equivalent to Newton's method. Further, we prove that, for all kernels, the mean shift procedure is a quadratic bound maximization.     

Mean shift denoising of point-sampled surfaces

This paper presents an anisotropic denoising/smoothing algorithm for point-sampled surfaces. Motivated by the impressive results of mean shift filtering on image denoising, we extend the concept to 3D surface smoothing by taking the vertex normal and the curvature as the range component and the vertex position as the spatial component. Then the local mode of each vertex on point-based surfaces is computed by a 3D mean shift procedure dependent on local neighborhoods that are adaptively obtained by a kdtree data structure. Clustering pieces of point-based surfaces of similar local mode provides a meaningful model segmentation. Based on the adaptively clustered neighbors, we finally apply a trilateral point filtering scheme that adjusts the position of sample points along their normal directions to successfully reduce noise from point-sampled surfaces while preserving geometric features.     

基于Mean shift算法的煤岩分界识别

提出了一种基于Mean shift算法的煤岩分界识别方案。首先介绍了Mean shift算法原理,通过关联图像的像素位置向量和灰度值构建了一个空间联合域;然后给出了适用于煤岩图像分割的带宽参数选择方法,以去除虚假孤立区域和杂散边界;最后利用煤岩图像的人造边界和真实边界进行仿真,结果表明Mean shift算法较K-means算法能更准确地获得煤岩分界线。     

Mean shift, mode seeking, and clustering

Mean shift, a simple interactive procedure that shifts each data point to the average of data points in its neighborhood is generalized and analyzed in the paper. This generalization makes some k-means like clustering algorithms its special cases. It is shown that mean shift is a mode-seeking process on the surface constructed with a “shadow” kernal. For Gaussian kernels, mean shift is a gradient mapping. Convergence is studied for mean shift iterations. Cluster analysis if treated as a deterministic problem of finding a fixed point of mean shift that characterizes the data. Applications in clustering and Hough transform are demonstrated. Mean shift is also considered as an evolutionary strategy that performs multistart global optimization     

Linear time-derivative trackers

The design of an ideal differentiator is a difficult and a challenging task. In this paper we discuss the properties and the limitations of two different structures of linear differentiation systems. The first time-derivative observer is formulated as a high-gain observer where the observer gain is calculated through a Lyapunov-like dynamical equation. The second one is given in Brunovski form in which the output to be differentiated appears as a control input and the differentiation gain is calculated from the dual Lyapunov equation of the first differentiation observer. A discrete-time version of the second form is given. Finally, illustrative examples are presented to show their strengths and weaknesses.     

基于分块背景更新的均值飘移跟踪算法

为提高融合背景信息均值飘移跟踪算法中的背景模型更新效率,提出基于分块背景更新的改进算法。将目标周围的背景区域划分成4块,分别以此4个子块的颜色直方图作为区域特征描述符进行计算,利用巴氏系数选择性地更新相似性较低的3个区域。在构建目标候选模型时,利用目标物体空间信息赋予候选目标区域像素点相应的距离权重。对整块背景信息更新策略和加权更新背景信息策略进行比较,比较结果表明,该算法具有更高的迭代收敛效率和更强的抗背景干扰能力。     

Mean shift based gradient vector flow for image segmentation

In recent years, gradient vector flow (GVF) based algorithms have been successfully used to segment a variety of 2-D and 3-D imagery. However, due to the compromise of internal and external energy forces within the resulting partial differential equations, these methods may lead to biased segmentation results. In this paper, we propose MSGVF, a mean shift based GVF segmentation algorithm that can successfully locate the correct borders. MSGVF is developed so that when the contour reaches equilibrium, the various forces resulting from the different energy terms are balanced. In addition, the smoothness constraint of image pixels is kept so that over- or under-segmentation can be reduced. Experimental results on publicly accessible datasets of dermoscopic and optic disc images demonstrate that the proposed method effectively detects the borders of the objects of interest.     

基于Mean shift的核窗宽自适应目标跟踪新算法

针对传统均值漂移算法(Mean shift)中核函数直方图对目标特征描述较弱、跟踪过程中核函数带宽的保持不变的缺点,提出了一种新的核函数带宽可变的Mean shift跟踪算法.在特定的色彩空间中,统计落入各区间的像素个数.并对各区间像素的位置建立高斯分布模型,采用二阶空间直方图实现目标建模,强化目标特征描述提高了跟踪的鲁棒性;结合边缘检测与角点检测选取目标特征点估算目标仿射模型确定伸缩尺度.适应目标多自由度变化下的跟踪.实验结果证明,该算法比原有算法跟踪效果更加准确和可靠.     

Mean shift: a robust approach toward feature space analysis

A general non-parametric technique is proposed for the analysis of a complex multimodal feature space and to delineate arbitrarily shaped clusters in it. The basic computational module of the technique is an old pattern recognition procedure: the mean shift. For discrete data, we prove the convergence of a recursive mean shift procedure to the nearest stationary point of the underlying density function and, thus, its utility in detecting the modes of the density. The relation of the mean shift procedure to the Nadaraya-Watson estimator from kernel regression and the robust M-estimators; of location is also established. Algorithms for two low-level vision tasks discontinuity-preserving smoothing and image segmentation - are described as applications. In these algorithms, the only user-set parameter is the resolution of the analysis, and either gray-level or color images are accepted as input. Extensive experimental results illustrate their excellent performance     

基于Mean shift和轨迹预测的目标跟踪算法

针对目标跟踪过程中出现的定位偏差问题,提出了Mean shift和轨迹预测相结合的运动目标跟踪算法.该算法首先根据目标已知位置信息采用最小二乘法拟合运动轨迹并得到预测位置,然后利用Mean shift算法得到目标最终位置.通过计算搜索误差判断是否发生严重遮挡情况,并给出相应处理策略.进行了一系列实验,验证了算法的有效性,并将实验结果与其他算法比较,表明该算法有效地提高了快速运动目标跟踪的精度,具有较强的鲁棒性.     

Mean shift tracker combined with online learning-based detector and Kalman filtering for real-time tracking

Color-based visual object tracking is one of the most commonly used tracking methods. Among many tracking methods, the mean shift tracker is used most often because it is simple to implement and consumes less computational time. However, mean shift trackers exhibit several limitations when used for long-term tracking. In challenging conditions that include occlusions, pose variations, scale changes, and illumination changes, the mean shift tracker does not work well. In this paper, an improved tracking algorithm based on a mean shift tracker is proposed to overcome the weaknesses of existing methods based on mean shift tracker. The main contributions of this paper are to integrate mean shift tracker with an online learning-based detector and to newly define the Kalman filter-based validation region for reducing computational burden of the detector. We combine the mean shift tracker with the online learning-based detector, and integrate the Kalman filter to develop a novel tracking algorithm. The proposed algorithm can reinitialize the target when it converges to a local minima and it can cope with scale changes, occlusions and appearance changes by using the online learning-based detector. It updates the target model for the tracker in order to ensure long-term tracking. Moreover, the validation region obtained by using the Kalman filter and the Mahalanobis distance is used in order to operate detector in real-time. Through a comparison against various mean shift tracker-based methods and other state-of-the-art methods on eight challenging video sequences, we demonstrate that the proposed algorithm is efficient and superior in terms of accuracy and speed. Hence, it is expected that the proposed method can be applied to various applications which need to detect and track an object in real-time.     

基于改进的Mean shift方法的图像特征提取

针对日益重要的图像信息和逐步发展的图像数据库技术,本文提出了一种基于改进的均值飘移算法和区域生长相结合的边缘检测方法提取图像的形状特征,并通过仿真试验证明其良好的有效性,为基于内容检索的图像检索提供了一个有效算法.     

基于梯度特征与彩色特征相融合的mean shift跟踪方法

针对mean shift跟踪方法中存在的光照变化不稳定问题,提出了基于梯度特征与彩色特征相融合的mean shift跟踪方法。首先分别提取目标的梯度特征和彩色特征,利用多尺度的相似度计算方法进行特征的匹配,然后通过最大化相似度对目标进行跟踪。通过物体和人体等运动目标的跟踪,验证了改进的跟踪算法在光照变化情况下的鲁棒性优于原有的算法,显著降低了跟踪位置误差。     

Dynamic response of electromagnetic spatial displacement trackers

Overall system latency--the elapsed time from input human motion until the immediate consequences of that input are available in the display--is one of the most frequently cited shortcoming of current virtual environment (VE) technology. Given that spatial displacement trackers are employed to monitor head and hand position and orientation in many VE applications, the dynamic response intrinsic to these devices is an unavoidable contributor to overall system latency. In this paper, we describe a testbed and method for measurement of tracker dynamic response that use a motorized rotary swing arm to sinusoidally displace the VE sensor at a number of frequencies spanning the bandwidth of volitional human movement. During the tests, actual swing arm angle and VE sensor reports are collected and time stamped. By calibrating the time stamping technique, the tracker's internal transduction and processing time are separated from data transfer and host computer software execution latencies. We have used this test-bed to examine several VE sensors--most recently to compare latency, gain, and noise characteristics of two commercially available electromagnetic trackers: Ascension Technology Corp.'s Flock of Birds(TM) and Polhemus Inc.'s Fastrak(TM).     

Error explanation with distance metrics

In the event that a system does not satisfy a specification, a model checker will typically automatically produce a counterexample trace that shows a particular instance of the undesirable behavior. Unfortunately, the important steps that follow the discovery of a counterexample are generally not automated. The user must first decide if the counterexample shows genuinely erroneous behavior or is an artifact of improper specification or abstraction. In the event that the error is real, there remains the difficult task of understanding the error well enough to isolate and modify the faulty aspects of the system. This paper describes a (semi-)automated approach for assisting users in understanding and isolating errors in ANSI C programs. The approach, derived from Lewis’ counterfactual approach to causality, is based on distance metrics for program executions. Experimental results show that the power of the model checking engine can be used to provide assistance in understanding errors and to isolate faulty portions of the source code.     

Bugzilla, ITracker, and other bug trackers

Bug-tracking helps the software developers in knowing what the error is, resolving it, and learning from it. Working on a software project includes managing the bugs we find. At first, we might list them on a spreadsheet. But when the number of bugs becomes too large and a lot of people must access and input data on them, we have to give up the spreadsheet and instead use a bug- or issue-tracking system. Many software projects reach this point, especially during testing and deployment when users tend to find an application's bugs. Nowadays we can choose among dozens of bug-tracking systems. This paper looks at two specific open source products and provides useful hints for working with any bug-tracking tool.     

Learning-based dynamic coupling of discrete and continuous trackers

We present a data-driven dynamic coupling between discrete and continuous methods for tracking objects of high dofs, which overcomes the limitations of previous techniques. In our approach, two trackers work in parallel, and the coupling between them is based on the tracking error. We use a model-based continuous method to achieve accurate results and, in cases of failure, we re-initialize the model using our discrete tracker. This method maintains the accuracy of a more tightly coupled system, while increasing its efficiency. At any given frame, our discrete tracker uses the current and several previous frames to search into a database for the best matching solution. For improved robustness, object configuration sequences, rather than single configurations, are stored in the database. We apply our framework to the problem of 3D hand tracking from image sequences and the discrimination between fingerspelling and continuous signs in American Sign Language.