Kernel-based Tracking Research Articles

Object contour tracking via adaptive data-driven kernel

We present a novel approach to non-rigid object tracking in this paper by deriving an adaptive data-driven kernel. In contrast with conventional kernel-based trackers which suffer from the constancy of kernel shape as well as scale and orientation selection problem when the tracking targets are changing in size, the adaptive kernel can robustly achieve the adaptation to target variation and act toward the actual target contour simultaneously with the mean shift iterations. Level set technique is novelly introduced to the mean shift sample space to both cope with insufficient low-level information and implement the adaptive kernel evolution and update. Since the active contour model is designed to drive the kernel constantly to the direction that maximizes the appearance similarity, this adaptive kernel can continually seize the target shape to give a better estimation bias and produce accurate shift of the mean. Finally, accurate target region can successfully avoid the performance loss stemmed from pollution of background pixels hiding inside the kernel and qualify the samples fed the next time step. Experimental results on a numer of challenging sequences validate the effectiveness of the technique.

Compact Scrutiny of Current Video Tracking System and its Associated Standard Approaches

With an increasing demands of video tracking systems with object detection over wide ranges of computer vision applications, it is necessary to understand the strengths and weaknesses of the present situation of approaches. However, there are various publications on different techniques in the visual tracking system associated with video surveillance application. It has been seen that there are prime classes of approaches that are only three, viz. point-based tracking, kernel-based tracking, and silhouette-based tracking. Therefore, this paper contributes to studying the literature published in the last decade to highlight the techniques obtained and brief the tracking performance yields. The paper also highlights the present research trend towards these three core approaches and significantly highlights the open-end research issues in these regards. The prime aim of this paper is to study all the prominent approaches of video tracking system which has been evolved till date in various literatures. The idea is to understand the strength and weakness associated with the standard approach so that new approaches could be effectively act as a guideline for constructing a new upcoming model. The prominent challenge in reviewing the existing approaches are that all the approaches are targeted towards achieving accuracy, whereas there are various other connected problems with internal process which has not been considered for e.g. feature extraction, processing time, dimensional problems, non-inclusion of contextual factor, which has been an outcome of the proposed review findings. The paper concluded by highlighting this as research gap acting as contribution of this review work and further states that there are some good possibilities of new work evolution if these issues are considered prior to developing any video tracking system. Overall, this paper offers an unbiased picture of the current state of video tracking approaches to be considered for developing any upcoming model.

A kernel-based trend pattern tracking system for portfolio optimization

We propose a novel kernel-based trend pattern tracking (KTPT) system for portfolio optimization. It includes a three-state price prediction scheme, which extracts both of the following and reverting patterns from the asset price trend to make future price predictions. Moreover, KTPT is equipped with a novel kernel-based tracking system to optimize the portfolio, so as to capture a potential growth of the asset price effectively. The kernel measures the similarity between the current portfolio and the predicted price relative to control the influence of each asset when optimizing the portfolio, which is different from some previous kernels that measure the probability of occurrence of a price relative. Extensive experiments on 5 benchmark datasets from real-world stock markets with various assets in different time periods indicate that KTPT outperforms other state-of-the-art strategies in cumulative wealth and other risk-adjusted metrics, showing its effectiveness in portfolio optimization.

Object tracking using both a kernel and a non-parametric active contour model

Abstract By combining both a kernel-based tracking and a non-parametric level set method, a novel framework for target tracking is proposed in this paper that robustly addresses tracking fast-moving and small targets with blurred edges. To establish our new framework, Kullback–Leibler divergence was adopted to measure the divergence between the foreground/background distributions and the target model, and the Bhattacharyya distance was adopted to measure the similarities between the foreground and background distributions. An image warping matrix is introduced into the framework to optimize the target function. The experimental results demonstrate the advantages of the proposed method compared with other methods.

Real time visual tracking using a spatially weighted von Mises mixture model

A real time, kernel based, model-free tracking algorithm is proposed, which employs a weighted von Mises mixture as the target’s appearance model. The mixture weights, which are provided by a spatial kernel, along with the hue values are used in order to estimate the parameters of the weighted von Mises mixture model. The von Mises distribution is suitable for circular data and it is employed in order to eliminate drawbacks in kernel-based tracking caused by eventual shifts of the target’s histogram bins. The weights allow a mean shift-like gradient based optimization by maximizing the weighted likelihood, which would not be feasible in the context of a standard von Mises mixture. Moreover, as only the hue component of the target is involved, many quantities of the algorithm may be pre-calculated for given parameters and therefore the algorithm can perform in real time, which is experimentally confirmed. Finally, it is shown that the proposed method has comparative performance in terms of accuracy and robustness with other state-of-the-art tracking algorithms.

Development of a tracking-based system for automated traffic data collection for roundabouts

Traffic data collection is essential for performance assessment, safety improvement and road planning. While automated traffic data collection for highways is relatively mature, that for roundabouts is more challenging due to more complex traffic scenes, data specifications and vehicle behavior. In this paper, the authors propose an automated traffic data collection system dedicated to roundabout scenes. The proposed system has mainly four steps of processing. First, camera calibration is performed for roundabout traffic scenes with a novel circle-based calibration algorithm. Second, the system uses enhanced Mixture of Gaussian algorithm with shaking removal for video segmentation, which can tolerate repeated camera displacements and background movements. Then, Kalman filtering, Kernel-based tracking and overlap-based optimization are employed to track vehicles while they are occluded and to derive the complete vehicle trajectories. The resulting vehicle trajectory of each individual vehicle gives the position, size, shape and speed of the vehicle at each time moment. Finally, a data mining algorithm is used to automatically extract the interested traffic data from the vehicle trajectories. The overall traffic data collection system has been implemented in software and runs on regular PC. The total processing time for a 3-hour video is currently 6 h. The automated traffic data collection system can significantly reduce cost and improve efficiency compared to manual data collection. The extracted traffic data have been compared to accurate manual measurements for 29 videos recorded on 29 different days, and an accuracy of more than 90% has been achieved.

Ground-Moving-Platform-Based Human Tracking Using Visual SLAM and Constrained Multiple Kernels

This paper proposes a robust ground-moving-platform-based human tracking system, which effectively integrates visual simultaneous localization and mapping (V-SLAM), human detection, ground plane estimation, and kernel-based tracking techniques. The proposed system systematically detects humans from recorded video frames of a moving camera and tracks the humans in the V-SLAM-inferred 3-D space via a tracking-by-detection scheme. To efficiently associate the detected human frame by frame, we propose a novel human tracking framework, combining the constrained-multiple-kernel tracking and the estimated 3-D information (depth), to globally optimize the data association between consecutive frames. By taking advantage of the appearance model and 3-D information, the proposed system not only achieves high effectiveness but also well handles occlusion in the tracking. Experimental results show the favorable performance of the proposed system, which efficiently tracks humans in a camera equipped on a ground-moving platform such as a dash camera and an unmanned ground vehicle.

Computational Acceleration of Real-Time Kernel-Based Tracking System

Object tracking in real-time is one of the applications of video processing, where the required computational cost is high due to intensive high data processing. In order to solve these problems, this paper presents an embedded solution, where the Hardware/Software (HW/SW) co-design architecture is used for the implementation of well-known kernel-based tracking system. In this algorithm, the target is searched in consecutive frame by maximizing the statistical match with similarity estimation of color distribution. The whole tracking system is implemented on low cost Field Programmable Gate Array (FPGA) device with image resolution of 1280[Formula: see text]720 pixels and target window size of 160[Formula: see text]80 pixels. The HW/SW co-design architecture is proposed to accelerate the computational speed of the system. The performance of the system is evaluated in terms of execution speed and frame rate compared with software based implementation. The hardware cost of design is also compared with other existing methods. The proposed design achieves 22 times computational speed and maximum 60 Frames Per Second (FPS) compared with software based design.

Robust Human Tracking Based on DPM Constrained Multiple-Kernel from a Moving Camera

In this paper, we attempt to solve the challenging task of precise and robust human tracking from a moving camera. We propose an innovative human tracking approach, which efficiently integrates the deformable part model (DPM) into multiple-kernel tracking from a moving camera. The proposed approach consists of a two-stage tracking procedure. For each frame, we first iteratively mean-shift several spatially weighted color histograms, called kernels, from the current frame to the next frame. Each kernel corresponds to a part model of a DPM-detected human. In the second step, conditioned on the tracking results of these kernels on the later frame, we then iteratively mean-shift the part models on that frame. The part models are represented by histogram of gradient (HOG) features, and the deformation cost of each part model provided by the trained DPM detector is used to constrain the movement of each detected body part from the first step. The proposed approach takes advantage of not only low computation owing to the kernel-based tracking, but also robustness of the DPM detector without the need of laborious human detection for each frame. Experimental results have shown that the proposed approach makes it possible to successfully track humans robustly with high accuracy under different scenarios from a moving camera.

Anomaly detection using a modified kernel-based tracking in the pantograph–catenary system

Condition monitoring is very important in railway systems to reduce maintenance costs and to increase the safety. A high power is needed for the movement of the electric train and collection of the current is critical. Faults occurred in the current collection system cause serious damage in the line and disrupt the railway traffic. When a wear occurs on the contact strip, the asymmetries and distortion are generated in supply voltage and current waveforms because of pantograph arcing. Therefore, the monitoring of pantograph–catenary system has been a hot topic in recent years. This paper deals with a method based on kernel-based object tracking for identifying the interaction between pantograph–catenary systems that gives useful information about the problems of catenary-pantograph systems. The method consists of two key components. The first component is based on the kernel based tracking of the contact wire. The contact point between pantograph and catenary is tracked and the obtained positions are saved as a signal. In the other hand, the foreground of each frame is found by using Gaussian mixture models (GMMs). The occurred arcs are detected by combining tracking and foreground detection methods. The second component employs S-transform for analyzing the pantograph problems, which are used to detect the faults occurred on pantograph strip. The experimental results imply that the proposed method is useful to detect burst of arcing, and irregular positioning of the contact wire.

A novel algebraic method for kernel-based object tracking

In the present paper, a new tracking method based on kernel tracking is proposed. The proposed method employs a novel algebraic algorithm to get the kernel movement. In contrast to the mean-shift method which uses a weighted kernel to reduce the effect of the background, the algebraic algorithm of the proposed method allows dividing the candidate area into two parts in order to identify the object and background regions. To detect the object and background regions, we propose measuring the similarity of weighted histogram for each part. The experiments show the superiority of the proposed method for the removal of the background. The effect of noise and background clutter is reduced by segmentation of the object which produces the narrow histogram. In conclusion, the ability of the proposed method for tracking in crowded and cluttered scenes is demonstrated.

Real-Time Boundary Detection of Fast Moving Object in Dark Surrounds

Dark surrounds make detection of moving target more difficult based on traditional methods. A real time identification of fast moving object under weak illumination is critical for some special applications. Traditional blob, contour and kernel-based tracking methods either need high computational loads or require normal illumination which limit their application. In this paper, we propose a new method trying to settle such difficulty based on temporal standard deviation. The performance of new method was evaluated with simulation data and real video data recorded by a simple imaging system. Combining hardware acceleration, a real time detection and visualization of fast moving boundary in dark environment can be achieved.

Dynamic Template Tracking and Recognition

In this paper we address the problem of tracking non-rigid objects whose local appearance and motion changes as a function of time. This class of objects includes dynamic textures such as steam, fire, smoke, water, etc., as well as articulated objects such as humans performing various actions. We model the temporal evolution of the object’s appearance/motion using a linear dynamical system. We learn such models from sample videos and use them as dynamic templates for tracking objects in novel videos. We pose the problem of tracking a dynamic non-rigid object in the current frame as a maximum a-posteriori estimate of the location of the object and the latent state of the dynamical system, given the current image features and the best estimate of the state in the previous frame. The advantage of our approach is that we can specify a-priori the type of texture to be tracked in the scene by using previously trained models for the dynamics of these textures. Our framework naturally generalizes common tracking methods such as SSD and kernel-based tracking from static templates to dynamic templates. We test our algorithm on synthetic as well as real examples of dynamic textures and show that our simple dynamics-based trackers perform at par if not better than the state-of-the-art. Since our approach is general and applicable to any image feature, we also apply it to the problem of human action tracking and build action-specific optical flow trackers that perform better than the state-of-the-art when tracking a human performing a particular action. Finally, since our approach is generative, we can use a-priori trained trackers for different texture or action classes to simultaneously track and recognize the texture or action in the video.

Mean Shift Tracking with Advanced Background-Weighted Histogram

Tracking objects in videos using mean shift technique has brought to public attention. In background-weighted histogram (BWH) algorithm proposed by Kernel-Based Object tracking attempts to reduce the interference of background target localization in mean shift tracking. However, in this paper we prove that the weights assigned to pixels in the target candidate region by BWH are proportional to those without background information, i.e. BWH does not introduce any new information because the mean shift iteration formula is invariant to the scale transformation of weights. We then propose an Advanced BWH (ABWH) formula by transforming only the target model but not the target candidate model. The ABWH scheme can effectively reduce background’s interference in target localization and for the target not well initialized, it can still robustly track the object, which is hard to achieve by the conventional target representation and it lead to faster convergence and more accurate localization than the usual target representation in mean shift tracking.

A Comparative Analysis of Kernel-Based Target Tracking Methods using Different Colour Feature Based Target Models

An effective target modeling is the root of a robust and efficient tracking system. Color feature is widely used feature space for target modeling in real time tracking applications because of its computational efficiency and invariance towards change in shape, scale and rotation. The effective use of this feature with kernel-based target tracking can lead to a robust tracking system. This paper provides a comparative analysis of the performance of three variants of kernel-based tracking system using color feature. The simulation results show that the target modeling using transformed background weighted target model will perform efficiently when initialized target has similar color feature with background while the combination of color-texture will be more accurate and robust when texture features are prominently present.

A Multiple Targets Appearance Tracker Based on Object Interaction Models

Kernel-based method has been proved to be effective in solving single-target tracking problem. However, facing more complicated multitarget tracking task, most classic kernel-based multitarget trackers do not model the interaction among targets successfully and simply track each target independently. Thus, they usually could not deal with “singularity” problem and fail in tracking the target when occlusions occur or distracters appear. Although multikernel methods may improve the performance by introducing more constraints, how to simulate the relationship and interaction among the tracked targets is still not fully investigated. In this paper, we discuss a very common scenario of multitarget tracking, in which a moving object's motion is not only determined by its virtual destination but also impacted by other neighboring objects. This phenomenon exists in many usual tracking applications such as human tracking, traffic monitoring, video surveillance, and so on, where an object usually moves toward a particular direction but meanwhile detours when close to others to avoid collision. Specifically, we propose a novel interaction model to explain the above phenomenon. Then by defining a new cost function, we embed this interaction model into a kernel-based tracker and further derive our interactive kernel-based multitarget tracker. Experimental results on various datasets demonstrate that our interaction model can alleviate “singularity” problem and, thus, the proposed tracking method could achieve superior performance in multitarget tracking.

Robust kernel-based tracking algorithm with background contrasting

The mean-shift algorithm has achieved considerable success in object tracking due to its simplicity and efficiency. Color histogram is a common feature in the description of an object. However, the kernel-based color histogram may not have the ability to discriminate the object from clutter background. To boost the discriminating ability of the feature, based on background contrasting, this letter presents an improved Bhattacharyya similarity metric for mean-shift tracking. Experiments show that the proposed tracker is more robust in relation to background clutter.

An Improved Adaptive Kernel-Based Object Tracking

Kernel-based density estimation technique, especially Mean-shift based tracking technique, is a successful application to target tracking, which has the characteristics such as with few parameters, robustness, and fast convergence. However, classic Mean-shift based tracking algorithm uses fixed kernel-bandwidth, which limits the performance when the target’s orientation and scale change. An Improved adaptive kernel-based object tracking is proposed, which extend 2-dimentional mean shift to 3-dimentional, meanwhile combine multiple scale theory into tracking algorithm. Such improvements can enable the algorithm not only track zooming objects, but also track rotating objects. The experimental results validate that the new algorithm can adapt to the changes of orientation and scale of the target effectively.

A method to track rotational motion for use in single-molecule biophysics

The double helical nature of DNA links many cellular processes such as DNA replication, transcription, and repair to rotational motion and the accumulation of torsional strain. Magnetic tweezers (MTs) are a single-molecule technique that enables the application of precisely calibrated stretching forces to nucleic acid tethers and to control their rotational motion. However, conventional magnetic tweezers do not directly monitor rotation or measure torque. Here, we describe a method to directly measure rotational motion of particles in MT. The method relies on attaching small, non-magnetic beads to the magnetic beads to act as fiducial markers for rotational tracking. CCD images of the beads are analyzed with a tracking algorithm specifically designed to minimize crosstalk between translational and rotational motion: first, the in-plane center position of the magnetic bead is determined with a kernel-based tracker, while subsequently the height and rotation angle of the bead are determined via correlation-based algorithms. Evaluation of the tracking algorithm using both simulated images and recorded images of surface-immobilized beads demonstrates a rotational resolution of 0.1°, while maintaining a translational resolution of 1-2 nm. Example traces of the rotational fluctuations exhibited by DNA-tethered beads confined in magnetic potentials of varying stiffness demonstrate the robustness of the method and the potential for simultaneous tracking of multiple beads. Our rotation tracking algorithm enables the extension of MTs to magnetic torque tweezers (MTT) to directly measure the torque in single molecules. In addition, we envision uses of the algorithm in a range of biophysical measurements, including further extensions of MT, tethered particle motion, and optical trapping measurements.

Carried Object Detection Using Ratio Histogram and its Application to Suspicious Event Analysis

This letter proposes a novel method to detect carried objects from videos and applies it for analysis of suspicious events. First of all, we propose a novel kernel-based tracking method for tracking each foreground object and further obtaining its trajectory. With the trajectory, a novel ratio histogram is then proposed for analyzing the interactions between the carried object and its owner. After color re-projection, different carried objects can be then accurately segmented from the background by taking advantages of Gaussian mixture models. After bag detection, an event analyzer is then designed to analyze various suspicious events from the videos. Even though there is no prior knowledge about the bag (such as shape or color), our proposed method still performs well to detect these suspicious events. As we know, due to the uncertainties of the shape and color of the bag, there is no automatic system that can analyze various suspicious events involving bags (such as robbery) without using any manual effort. However, by taking advantages of our proposed ratio histogram, different carried bags can be well segmented from videos and applied for event analysis. Experimental results have proved that the proposed method is robust, accurate, and powerful in carried object detection and suspicious event analysis.