Proposed Particle Filter Research Articles

An evolutionary particle filter with the immune genetic algorithm for intelligent video target tracking

Particle filter algorithm is widely used for target tracking using video sequences, which is of great importance for intelligent surveillance applications. However, there is still much room for improvement, e.g. the so-called “sample impoverishment”. It is brought by re-sampling which aims to avoid particle degradation, and thus becomes the inherent shortcoming of the particle filter. In order to solve the problem of sample impoverishment, increase the number of meaningful particles and ensure the diversity of the particle set, an evolutionary particle filter with the immune genetic algorithm (IGA) for target tracking is proposed by adding IGA in front of the re-sampling process to increase particle diversity. Particles are regarded as the antibodies of the immune system, and the state of target being tracked is regarded as the external invading antigen. With the crossover and mutation process, the immune system produces a large number of new antibodies (particles), and thus the new particles can better approximate the true state by exploiting new areas. Regulatory mechanisms of antibodies, such as promotion and suppression, ensure the diversity of the particle set. In the proposed algorithm, the particle set optimized by IGA can better express the true state of the target, and the number of meaningful particles can be increased significantly. The effectiveness and robustness of the proposed particle filter are verified by target tracking experiments. Simulation results show that the proposed particle filter is better than the standard one in particle diversity and efficiency. The proposed algorithm can easily be extended to multiple objects tracking problems with occlusions.

Specified Object Tracking Problem in an Environment of Multiple Moving Objects

Video based object tracking normally deals with non-stationary image streams that change over time. Robust and real time moving object tracking is considered to be a problematic issue in computer vision. Multiple object tracking has many practical applications in scene analysis for automated surveillance. In this paper, we introduce a specified object tracking based particle filter used in an environment of multiple moving objects. A differential image region based tracking method for the detection of multiple moving objects is used. In order to ensure accurate object detection in an unconstrained environment, a background image update method is used. In addition, there exist problems in tracking a particular object through a video sequence, which cannot rely only on image processing techniques. For this, a probabilistic framework is used. Our proposed particle filter has been proved to be robust in dealing with nonlinear and non-Gaussian problems. The particle filter provides a robust object tracking framework under ambiguity conditions and greatly improves the estimation accuracy for complicated tracking problems.

A new state estimation method for chaotic signals: Map-particle filter method

Over the past few decades, research has been conducted regarding the applicability of chaotic behavior in consumer and industrial applications, including secure communication. In this paper, we develop a new particle filter termed as MAP-particle filter and apply it to a secure communication. The proposed particle filter estimates not only the chaotic state but also the secure message, thereby improving the secure communication performance. A computer simulation is conducted to demonstrate the effectiveness of our proposed method.

Inferring Parameters of Gene Regulatory Networks via Particle Filtering

Gene regulatory networks are highly complex dynamical systems comprising biomolecular components which interact with each other and through those interactions determine gene expression levels, that is, determine the rate of gene transcription. In this paper, a particle filter with Markov Chain Monte Carlo move step is employed for the estimation of reaction rate constants in gene regulatory networks modeled by chemical Langevin equations. Simulation studies demonstrate that the proposed technique outperforms previously considered methods while being computationally more efficient. Dynamic behavior of gene regulatory networks averaged over a large number of cells can be modeled by ordinary differential equations. For this scenario, we compute an approximation to the Cramer-Rao lower bound on the mean-square error of estimating reaction rates and demonstrate that, when the number of unknown parameters is small, the proposed particle filter can be nearly optimal.

SU‐GG‐J‐18: Robust Tracking of Fiducials and Interventional Tools during Image‐Guided Interventions

Purpose: The purpose of image‐guidance during clinical interventions is to provide information on the location of objects of interest, often relative to other anatomical structures. Robust methods for real‐time tracking of such objects are necessary to provide quantitative information to the interventionalist. Most tracking algorithms fail when the assumptions of the algorithm are not met or when cases are encountered that are unaccounted for by the model used. This study highlights some of the challenges involved in tracking objects in 2D image sequences, particularly in X‐ray fluoroscopy, and proposes methods to address them. Method and Materials: Experiments are conducted by simulating three fundamental cases of noise, occlusion and unexpected dynamics in X‐ray fluoroscopic images. An extension to particle filters is developed and is used to track objects under these challenging conditions. The proposed extension to particle filters (i) adapts the number of particles used during tracking in response to noise in the image, (ii) modulates the resolution of particles for robustness to occlusion, and (iii) adjusts to unexpected dynamics or ‘surprises’ in motion exhibited by the object of interest. The performance of this algorithm is contrasted with a conventional particle filter using the mean squared error (M.S.E.) as a metric. Results: When compared to a conventional particle filter, a reduction in the M.S.E is observed for the proposed particle filter in all three fundamental cases evaluated. Conclusion: The proposed adaptive particle filter offers an improvement in robustness to factors unaccounted for in the dynamic and observation models that constitute the tracking algorithm. This method would be valuable in image‐guided interventions as it can provide real‐time feedback that can improve the performance of a surgeon, control robotics, and optimize imaging parameters.

Contour segmentation in 2D ultrasound medical images with particle filtering

Object segmentation in medical images is an actively investigated research area. Segmentation techniques are a valuable tool in medical diagnostics for cancer tumours and cysts, for planning surgery operations and other medical treatment. In this paper, a Monte Carlo algorithm for extracting lesion contours in ultrasound medical images is proposed. An efficient multiple model particle filter for progressive contour growing (tracking) from a starting point is developed, accounting for convex, non-circular forms of delineated contour areas. The driving idea of the proposed particle filter consists in the incorporation of different image intensity inside and outside the contour into the filter likelihood function. The filter employs image intensity gradients as measurements and requires information about four manually selected points: a seed point, a starting point, arbitrarily selected on the contour, and two additional points, bounding the measurement formation area around the contour. The filter performance is studied by segmenting contours from a number of real and simulated ultrasound medical images. Accurate contour segmentation is achieved with the proposed approach in ultrasound images with a high level of speckle noise.

FRAGMENTED PARTICLES-BASED ROBUST OBJECT TRACKING WITH FEATURE FUSION

Tracking people or objects across multiple cameras and maintaining a track within a camera is a challenging task in applications such as video surveillance, traffic monitoring, event detection and recognition. Some of the major challenges while tracking a target are illumination/scale changes and partial occlusion. In this paper, we propose a novel tracking framework using particle filter to efficiently track an object within a camera and a blob-based target association scheme for tracking across cameras. The proposed particle filter tracking algorithm uses a fragment-based approach to model the target and track it by fusing color and gradient features. The proposed solution incorporates coarser level spatial information by fragmenting each particle and is shown to be useful for tracking under partial occlusion. A fast yet robust model update is employed to overcome illumination changes. Experimental results show (i) the robustness of the fragment-based tracking approach with respect to illumination/scale change and partial occlusion and (ii) tracking persons across two cameras.

A New Particle Filter for Nonlinear Filtering Problems

Particle filters have gained special attention of researchers in various fields.The key idea of this technique is to represent the posterior density by sets of weighed samples.This paper proposes a new particle filter which is based on the extended Kalman filter and the Unscented Kalman filter.It first uses the former to generate an estimate of the state at time k,and then uses the latter to repeat the process and to gain the final estimate of the state and corresponding covariance at time k.In the experiments,the authors test five different particle filters on two different nonlinear systems.The experimental results indicate that the proposed particle filter has much better performance than the other four particle filters do.

Dynamic View Planning by Effective Particles for Three-Dimensional Tracking

In this paper, we propose a new approach to dynamically manage the viewpoint of a vision system for optimal 3-D tracking using particle techniques. We adopt the effective sample size in the proposed particle filter as a criterion for evaluating tracking performance and employ it to guide the view-planning process for finding the best viewpoint configuration. In our approach, the vision system is designed and configured to achieve the largest number of effective particles, which minimizes tracking error by revealing the system to a better swarm of importance samples and interpreting posterior states in a better way. Superiorities of our method are shown by comparison with the resampling particle filter and other view-planning methods.

A particle filter approach for tomographic imaging based on different state-space representations

In electrical capacitance tomography (ECT) the main focus is on the reconstruction of distinct objects with sharp transitions between different phases. Being inherently ill-posed, the reconstruction algorithm requires some sort of regularization to stabilize the solution of the inverse problem. However, introducing regularization may counteract the reconstruction of well-defined contours for grid-based methods. Level set propagation approaches which also rely on regularization are able to model sharp phase boundaries but suffer from high computational demands. In this contribution, two different state-space representations of closed contours based on B-splines and on Fourier descriptors are investigated. Both approaches allow us to describe the problem with only a small set of state-space variables. Regularization is incorporated implicitly which can be directly interpreted in the object domain as it relates to smooth contours. To solve the inverse problem, statistical inversion is performed by means of particle filtering providing the opportunity to conveniently incorporate prior information and to take measurement uncertainties into account. The proposed particle filter approach is compared to an extended Kalman filter realization in terms of complexity, computation time and estimation accuracy.

Particle Filter Design Using Importance Sampling for Acoustic Source Localisation and Tracking in Reverberant Environments

Sequential Monte Carlo methods have been recently proposed to deal with the problem of acoustic source localisation and tracking using an array of microphones. Previous implementations make use of the basic bootstrap particle filter, whereas a more general approach involves the concept of importance sampling. In this paper, we develop a new particle filter for acoustic source localisation using importance sampling, and compare its tracking ability with that of a bootstrap algorithm proposed previously in the literature. Experimental results obtained with simulated reverberant samples and real audio recordings demonstrate that the new algorithm is more suitable for practical applications due to its reinitialisation capabilities, despite showing a slightly lower average tracking accuracy. A real-time implementation of the algorithm also shows that the proposed particle filter can reliably track a person talking in real reverberant rooms.

The hybrid annealed particle filter

In this paper, a new particle filter based on sequential importance sampling (SIS) is proposed for the on-line estimation of non-Gaussian nonlinear systems. In this filtering method, state parameters separation and an annealing parameter are used to produce importance function. Since the distribution function makes full use of the prior, likelihood, and statistical characteristics of noise and the newest observation data, it is much closer to posterior distributions. Theoretical analysis and simulation show that the performance of proposed particle filter outperforms the standard particle filter and the extended Kalman filter.