Abstract:Siamese networks have achieved great success in object tracking due to the balance of precision and speed. However, Siamese trackers usually utilize the local feature of the last layer, which may degrade tracking performance in some difficult scenarios. In this paper, we propose a novel Channel and Spatial Attention-guided Residual learning framework for Tracking, referred to as CSART, which can improve feature representation of Siamese networks by exploiting self-attention mechanism to capture powerful contextual information. Specifically, to be efficient and seamless integration, different kinds of self-attention are appended on the template and search branches of Siamese networks respectively, that model global semantic inter-dependencies in channel and spatial dimensions. To avoid representation degradation, we consider to adaptively aggregate basic feature and its attention-weighted features with residual learning. Furthermore, a joint loss consisting of classic logistic loss as well as focal softmax loss is designed to emphasize difficult samples and guide the learning process of the whole model. Benefiting from the above scheme, CSART alleviates the over-fitting problem to some extent and enhances the discriminability. Extensive experiments on six popular tracking datasets indicate that the proposed tracker achieves better performance with a speed of 65 fps than other state-of-the-art trackers. (c) 2020 Elsevier B.V. All rights reserved. Siamese networks have achieved great success in object tracking due to the balance of precision and speed. However, Siamese trackers usually utilize the local feature of the last layer, which may degrade tracking performance in some difficult scenarios. In this paper, we propose a novel Channel and Spatial Attention-guided Residual learning framework for Tracking, referred to as CSART, which can improve feature representation of Siamese networks by exploiting self-attention mechanism to capture powerful contextual information. Specifically, to be efficient and seamless integration, different kinds of self-attention are appended on the template and search branches of Siamese networks respectively, that model global semantic inter-dependencies in channel and spatial dimensions. To avoid representation degradation, we consider to adaptively aggregate basic feature and its attention-weighted features with residual learning. Furthermore, a joint loss consisting of classic logistic loss as well as focal softmax loss is designed to emphasize difficult samples and guide the learning process of the whole model. Benefiting from the above scheme, CSART alleviates the over-fitting problem to some extent and enhances the discriminability. Extensive experiments on six popular tracking datasets indicate that the proposed tracker achieves better performance with a speed of 65 fps than other state-of-the-art trackers.

A Robust Tracking Architecture Using Tracking Failure Detection in Siamese Trackers.

AFAT: Adaptive Failure-Aware Tracker for Robust Visual Object Tracking

ASiam: Adaptive Siamese Regression Tracking with Adversarial Template Generation and Motion-Based Failure Recovery.

Siamese Tracking Network with Spatial-Semantic-Aware Attention and Flexible Spatiotemporal Constraint

Motion Guided Siamese Trackers for Visual Tracking

Improved Fully Convolutional Siamese Networks for Visual Object Tracking Based on Response Behaviour Analysis.

Long Short-Term Memory Improved Siamese Network for Robust Target Tracking.

Learning Motion-Perceive Siamese network for robust visual object tracking

Discriminative and Robust Online Learning for Siamese Visual Tracking

Siamese Residual Network for Efficient Visual Tracking

An Adaptive Kalman-Correlation Based Siamese Network Tracker for Visual Object Tracking

SiamRDT: An Object Tracking Algorithm Based on a Reliable Dynamic Template

CSART: Channel and Spatial Attention-Guided Residual Learning for Real-Time Object Tracking.

Unsupervised Learning of Accurate Siamese Tracking

Long-term object tracking based on joint tracking and detection strategy with Siamese network

An object tracking framework with recapture based on correlation filters and Siamese networks

Toward Robust Visual Object Tracking With Independent Target-Agnostic Detection and Effective Siamese Cross-Task Interaction

Object Tracking Algorithm Based on Channel-interconnection-spatial Attention Mechanism and Siamese Region Proposal Network

Adaptive distractor-aware for siamese tracking via enhancement confidence evaluator

Siamada: Visual Tracking Based on Siamese Adaptive Learning Network

Learning Adaptive Updating Siamese Network for Visual Tracking