Abstract:Visual tracking is a challenging task in computer vision. Recently correlation filter based trackers have gained much attention due to their high efficiency and impressive performance. Several methods have been developed to utilize deep features extracted from Convolutional Neural Networks (CNNs) for correlation filter tracking. Despite their success, most of these approaches suffer from low tracking speed due to high computation burden of the deep models. In this paper, we propose a CNN based real-time deep tracker for robust visual tracking. We first exploit hierarchical deep features as target representation, which can better distinguish the target from the backgrounds in the presence of complex situations. Before applied to learn correlation filters, the channel number of hierarchical features is reduced by the principal component analysis (PCA). This method can decrease both feature redundancy and computation. Then we construct a reliable map with features from the last convolutional layer as they encode the most semantic information. The reliable map is used to constraint the searching area where the target most likely exist. To further handle long-term tracking, we improve the model discrimination capability by introducing the original template into the current correlation filter model. As the target movement is usually small between two adjacent frames, we reuse the deep features from location step to update current model. The features are reused by making a circular shift with position displacement of the target, which increases the tracking speed significantly. Extensive experimental results on large benchmarks show that our proposed tracker can perform at real-time and achieves the state-of-the-art performance.

A Multi-Scale Kernel Correlation Filter Tracker with Feature Integration and Robust Model Updater

A Scale Adaptive Kernel Correlation Filter Tracker With Feature Integration

High-speed Tracking with Multi-Templates Correlation Filters

Scale Adaptive Kernelized Correlation Filter Tracker with Feature Fusion

Kernel Correlation Filters for Visual Tracking with Adaptive Fusion of Heterogeneous Cues

A novel adaptive kernel correlation filter tracker with multiple feature integration

Visual Tracking Based on Multi-Feature and Fast Scale Adaptive Kernelized Correlation Filter.

A Scale Adaptive Kernel Correlation Filter-Based Tracker with Optimized Update Strategy

Robust Kernel Correlation Filter Tracker Based on Adaptive Scale and Occlusion Detection

Deep Correlation Filter based Real-Time Tracker

Robust Scale Adaptive and Real-Time Visual Tracking with Correlation Filters

Multi-Scale Correlation Filtering Tracker Based on Adaptive Feature Selection

Correlation Filter Tracking Via Bootstrap Learning

A Robust Correlation Filter Based Tracker with Rich Representation and a Relocation Component.

A Background-Aware Tracker with Multi-Features Correlation Filters.

A Weight-adaptive Algorithm of Multi Feature Fusion Based on Kernel Correlation Filtering for Target Tracking.

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

Correlation Filters With Adaptive Memories And Fusion For Visual Tracking

Aggregating Correlation Filter with Multiple Kernels Learning for Robust Visual Object Tracking.

A Multiple Feature Fused Model for Visual Object Tracking Via Correlation Filters

Real-Time Object Tracking Via Adaptive Correlation Filters