Abstract:The problem of serious air pollution such as haze and fog and other complex scenes has posed great challenges for target tracking in computer vision. Single-feature approaches demonstrating unstable performances lead to poor tracking. Through the multi-feature fusion and the focus of attention (FOA) mechanism in biological vision system, these problems under complex scenes can be solved. Accordingly, a tracking algorithm based on multi-feature fusion and spatio-context correlation is proposed. Accurate expression of multi-feature fusion by color, texture, and edge features, and adaptive weighted update by information entropy, has greatly enhanced the adaptability to environment variations. Then, combined with the spatio-temporal context algorithm, the target can be achieved accurately. Compared with the state-of-the-art tracking algorithms, the experiment results validate the effectiveness of our method in hazy scenes, furthermore, our method also results in the promotion of image vision quality. Specifically, the average center error is reduced to 0.9440 pixel, and the average overlap rate and the FPS rise to 0.8700 and 4.4302, respectively. Additionally, the halo artifacts and color cast of restored images have been avoided by the popular dehazing algorithm. Our proposed tracking algorithm outperforms the state-of-the-art methods in accuracy, robustness, and real-time even in complex real-world hazy scenes.

Robust Tracking Algorithm Based on Multi-Feature Fusion and Particle Filter

Object Tracking Model Based on Fusing Multiple Weighted Distribution Field Features

Robust Visual Tracking Base on Adaptively Multi-Feature Fusion and Particle Filter

The improved particle filter for object tracking

Robust Visual Tracking Based on Joint Multi-Feature Histogram by Integrating Particle Filter and Mean Shift

Robust Object Tracking With Multi-Feature Selection and Particle Filter

Particle Filter Based Visual Tracking with Multi-Cue Adaptive Fusion

Adaptive Tracking Algorithm Based on Particle Grouped Optimization

Robust Object Tracking Combining Color and Scale Invariant Features

Object contour tracking using multi-feature fusion based particle filter

Hierarchical particle filter tracking algorithm based on multi-feature fusion

Particle Filtering Algorithm Based on Dynamic Multi-Feature Fusion

Dim and Small Target Tracking Using an Improved Particle Filter Based on Adaptive Feature Fusion

Anti-occlusion Particle Filter Object-Tracking Method Based on Feature Fusion

Object Tracking Based on Multi-Information and Auto-Adapted Particle Filter

Visual tracking using particle filters based on data fusion

Spatio-Context-Based Target Tracking with Adaptive Multi-Feature Fusion for Real-World Hazy Scenes.

Robust appearance-guided particle filter for object tracking with occlusion analysis

Visual Tracking Via Adaptive Interactive Fusion

Appearance Model-Based Particle Filter for Robust Tracking

Robust Fragments-Based Tracking with Adaptive Feature Selection