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.

Human-Following of Mobile Robots Based on Object Tracking and Depth Vision

A Skeleton and Visual Tracking Fusion Based Person-Following System for Mobile Service Robots

An Efficient Human-Following Method by Fusing Kernelized Correlation Filter and Depth Information for Mobile Robot.

Accurate and Real-Time 3-D Tracking for the Following Robots by Fusing Vision and Ultrasonar Information

Human Tracking Using Improved Sample-Based Joint Probabilistic Data Association Filter.

Human Tracking and Following in Dynamic Environment for Service Robots

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

Long-Term Real-Time Correlation Filter Tracker For Mobile Robot

Monocular Vision Tracking Based on Hybrid Particle Filters for a Person Following Robot

Human Following for Wheeled Robot with Monocular Pan-tilt Camera

Human Following for Outdoor Mobile Robots Based on Point‐Cloud's Appearance Model

Mobile Robot Tracking Method Based on Improved YOLOv8 Pedestrian Detection Algorithm

Vision-based object tracking method of mobile robot

Deep Correlation Filter based Real-Time Tracker

A Cascade Tracking Algorithm Based On Artificial Beacon And Hog Features

Person-following for Telepresence Robots Using Web Cameras

A Monocular Vision-Based Human-Following Approach for Mobile Robots

A Real-time Moving Target Following Mobile Robot System with Depth Camera

Sensor Integration for Person Tracking and Following with Mobile Robot

A Robust Person Tracking and Following Approach for Mobile Robot

Person following of binocular robot by extracting multiple features