Abstract:Visual tracking is a critical task in many computer vision applications such as surveillance, vehicle tracking, and motion analysis. The challenges in designing a robust visual tracking algorithm are caused by the presence of background clutter, occlusion, and illumination changes. In this paper, we propose a visual tracking algorithm in a particle filter framework to overcome these three challenging issues. Particle filter is an inference technique for estimating the unknown motion state from a noisy collection of observations, so we employ particle filter to learn the trajectory of a target. The proposed algorithm depends on the learned trajectory to predict the position of a target at a new frame, and corrects the predication by a process that can be entitled field transition. At the beginning of the tracking stage, a set of disturbance templates around the target template are accurately selected and defined as particles. During tracking, a position of the tracked target is firstly predicted based on the learned motion state, and then we take the normalized cross-correlation coefficient as a level to select the most suitable field transition parameters of the predicted position from the corresponding parameters of the particles. After judging the target is not occluded, we apply the named field transition with the selected parameters to compensate the predicted position to the accurate location of the target, meanwhile, we make use of the calculated cross-correlation coefficient as a posterior knowledge to update the weights of all the particles for the next prediction. In order to evaluate the performance of the proposed tracking algorithm, we test the approach on challenging sequences involving heavy background clutter, severe occlusions, and drastic illumination changes. Comparative experiments have demonstrated that this method makes a more significant improvement in efficiency and accuracy than two previously proposed algorithms: the mean shift tracking algorithm (MS) and the covariance tracking algorithm (CT).

Biologically Inspired Particle Filter for Robust Visual Tracking1

Visual tracking using the kernel based particle filter and color distribution

Enhancing Particle Swarm Optimization Based Particle Filter Tracker

The improved particle filter for object tracking

Visual Tracking Using an Insect Vision Embedded Particle Filter

Adaptive Visual Tracking Using Particle Filter

Insect vision inspired particle filter for visual tracking

Better Proposal Distributions: Object Tracking Using Unscented Particle Filter

Augmented Particle Filtering for Efficient Visual Tracking

Multifeature Fusion Tracking in a Particle Filter Framework

Visual tracking based on harmony search particle filter

Real-time Particle Filter Tracking Method Based on Human Vision Non-uniform Characteristics

Adaptive Particle Filter for Robust Visual Tracking

A Robust Particle Filter for People Tracking

Efficient Particle Filter and Its Application in Contour Tracking

Particle Filter Based on Color Feature with Contour Information Adaptively Integrated for Object Tracking

Top-Down Visual Attention Integrated Particle Filter for Robust Object Tracking

Visual Tracking with adaptive layered-optimizing particles in Multifeature Particle Filtering Framework

An Adaptive Unscented Particle Filter Tracking Algorithm Based On Color Distribution And Wavelet Moment

Adaptive Probabilistic Tracking with Visual Saliency Selection Reliable Particles

A Field Transition Particle Filter Tracking Algorithm