Abstract:Correlation filter-based video object-tracking algorithms have gained widespread attention due to their efficiency and excellent tracking performance. However, traditional correlation filtering tracking algorithms possess several limitations. (1) They extract image features only by using rule sampling, which ignores the shape information of the target, resulting in insufficient discriminative power of the features. (2) They focus only on the difference between the background and the object, ignoring the effects of more challenging intra-class interference. (3) They do not further evaluate the reliability of the optimal candidate samples, resulting in easy failure in occlusion scenarios. This paper proposes an asymmetric background-aware correlation filter method for object tracking with deep features to solve the above limitations. First, an asymmetrical background-aware sampling method based on the shape information of the object is proposed. This sampling method significantly differs from the symmetrical sampling method in the traditional correlation filter framework. By exploring the shape information of the object, improving the otherness between the background and object samples is easy, thus suppressing the intra-class interferences. Second, deep neural networks are introduced in the correlation filter framework to extract the deep object features and a spatio-temporal regularization factor is adaptively assigned to suppress the boundary effects, intra-class distractors and aberrance between frames. Finally, a multi-modal object pool is constructed to evaluate the optimal candidate sample in each frame. This template pool fully exploits object diversity and solves the tracking drift and failure caused by invalid appearance changes in scenes such as occlusion and vigorous motion scenarios. To validate the effectiveness of the proposed method, it was compared with the state-of-the-art methods on public datasets. The experimental results show that the tracking performance of the proposed method is competitive.

An Improved Kernel Correlation Filter for Occlusion Target Tracking

Long Time Target Tracking Based on Kernel Correlation Filtering

High-speed Tracking with Multi-Templates Correlation Filters

Long-Term Object Tracking Algorithm Based on Kernelized Correlation Filter

RETRACTED: Multiscale fast correlation filtering tracking algorithm based on a feature fusion model (Retracted article. See vol. 34, 2022)

An improved target tracking method based on extraction of corner points

A Scale Adaptive Kernel Correlation Filter Tracker With Feature Integration

A novel kernelized correlation filter by fusing multiple feature response maps, enhanced target re-detection, and improved model updating for visual tracking

Robust Visual Tracking via Adaptive Kernelized Correlation Filter

Improving the Anti-Occlusion Ability of Correlation Filter-Based Trackers Via Segmentation

Multiple Object Tracking with Kernelized Correlation Filters in Urban Mixed Traffic

Infrared Small Target Tracking Based on Target Spatial Distribution With Improved Kernelized Correlation Filtering

HSKP-CF Algorithm Based on Target Tracking for Mobile Following Robot

Extended Kernel Correlation Filter for Abrupt Motion Tracking.

Learning Local–Global Multiple Correlation Filters for Robust Visual Tracking with Kalman Filter Redetection

Real-Time Scalable Visual Tracking Via Quadrangle Kernelized Correlation Filters.

Correlation filter tracking algorithm based on spatial-temporal regularization and context awareness

Parallel Correlation Filters for Real-Time Visual Tracking

Adaptive Distraction Context Aware Tracking Based on Correlation Filter

Collaborative model tracking with robust occlusion handling

Deep-Feature-Based Asymmetrical Background-Aware Correlation Filter for Object Tracking