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.

Robust Visual Tracking with Deep Feature Fusion

Multi-features Guided Robust Visual Tracking.

Deep Correlation Filter based Real-Time Tracker

Learning Fully Convolutional Network for Visual Tracking with Multi-Layer Feature Fusion

Robust Visual Tracking Based on Fusional Multi-Correlation-Filters with a High-Confidence Judgement Mechanism

Exploiting multi-scale hierarchical feature representation for visual tracking

Adaptive feature fusion for visual object tracking

Robust Joint Discriminative Feature Learning for Visual Tracking

Visual Tracking Based on Multi-Feature Fusion

Robust Visual Tracking Method via Deep Learning

Siamese Network Based Features Fusion For Adaptive Visual Tracking

A Robust Tracking with Low-Dimensional Target-Specific Feature Extraction.

Real-time tracking based on deep feature fusion

Visual Tracking with Kernelized Correlation Filters Based on Multiple Features

Adaptive Features Fusion Correlation Filter for Real-time Object Tracking

Robust Visual Object Tracking Based on Feature Channel Weighting and Game Theory

HCDC-SRCF tracker: Learning an adaptively multi-feature fuse tracker in spatial regularized correlation filters framework

Robust Visual Tracking Via Collaborative and Reinforced Convolutional Feature Learning

Long-Term Tracking Based on Multi-Feature Adaptive Fusion for Video Target

Learning Compact Target-Oriented Feature Representations for Visual Tracking

Robust Visual Track Using an Ensemble Cascade of Convolutional Neural Networks