Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jun 25, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

Adaptive sparse attention-based compact transformer for object tracking.

Fei Pan¹, Lianyu Zhao¹, Chenglin Wang²

¹School of Computer Science and Engineering, Tianjin University of Technology, Liqizhaung street, Tianjin, 300384, China.

Scientific Reports

|May 28, 2024

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Antibacterial, In Vitro Anti-Inflammatory and Anti-Acne Activities of <i>Patchouli</i> Essential Oil.

Pharmaceuticals (Basel, Switzerland)·2026

Same author

Single-cell analyses identify the ginseng embryonic protoderm as a native compartment for high-efficiency ginsenoside production.

Nature communications·2026

Same author

[Analysis of complicating factors in endodontic microsurgery].

Hua xi kou qiang yi xue za zhi = Huaxi kouqiang yixue zazhi = West China journal of stomatology·2026

Same author

Impact of aerosol acidity on the heterogeneous chemistry of halogenated phenols.

Environmental pollution (Barking, Essex : 1987)·2026

Same author

Network toxicology and molecular docking identify BRCA1 as a functional target of the dietary carcinogen PhIP in colorectal cancer.

Discover oncology·2026

Same author

Lysosome-related biomarkers in peripheral blood immune cells discriminate sepsis from SIRS.

Scientific reports·2026

Same journal

Integrated multi-assessment and structural performance index framework for stacking-sequence optimisation of natural fibre reinforced laminates.

Scientific reports·2026

Same journal

SuperiorGAT: graph attention networks for sparse LiDAR point cloud reconstruction in autonomous systems.

Scientific reports·2026

Same journal

The effect of stretching the pectoralis major, sternocleidomastoid, and iliopsoas muscles on 800 m swimming performance in master swimmers.

Scientific reports·2026

Same journal

ISNR-PQC: isometry noise resilience post quantum cryptography primitive.

Scientific reports·2026

Same journal

Identification of high-yielding and stable genotypes of barley in the cold climate of Iran using AMMI and GGE biplot models.

Scientific reports·2026

Same journal

Bayesian negative binomial modelling of spatial and temporal patterns of road traffic deaths in Ghana.

Scientific reports·2026

See all related articles

This study introduces ASACTT, a novel object tracking method that enhances global information capture and target-specific attention. ASACTT achieves state-of-the-art performance with improved efficiency and robustness in object tracking.

Area of Science:

Computer Vision
Artificial Intelligence
Machine Learning

Background:

Transformer-based Siamese networks are effective for object tracking but struggle with global information capture and feature representation due to ResNet backbones.
Existing methods face challenges in attending to target-relevant information using multi-head self-attention (MSA) and exhibit robustness issues and high model complexity during online tracking.

Purpose of the Study:

To address the limitations of current object tracking methods by proposing a novel tracker, ASACTT.
To enhance global information extraction, improve target-specific attention, and ensure robust tracking with adaptive appearance variations.

Main Methods:

Improved Swin-Transformer-Tiny backbone for enhanced global information extraction.
Adaptive Sparse Attention (ASA) mechanism to focus on target-specific details within the search region.

Keywords:

Adaptive sparse attention Object tracking Siamese network Transformer

More Related Videos

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

Methods to Test Visual Attention Online

Methods to Test Visual Attention Online

Published on: February 19, 2015

Related Experiment Videos

Last Updated: Jun 25, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images

Published on: April 21, 2023

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

Methods to Test Visual Attention Online

Methods to Test Visual Attention Online

Published on: February 19, 2015

Dynamic Template Updater (DTU) using position encoding and historical data for adaptive appearance tracking and reduced complexity.

Main Results:

ASACTT demonstrates high comparability to state-of-the-art methods across five benchmark datasets.
Achieved an outstanding success score of 75.3% at 36 FPS on the GOT-10K dataset.
Significantly surpassed other trackers with comparable model parameters, indicating improved accuracy and efficiency.

Conclusions:

The proposed ASACTT tracker effectively overcomes limitations in global information capture and target attention in Transformer-based Siamese networks.
ASACTT offers a robust and efficient solution for object tracking, balancing accuracy with reduced model complexity.
The adaptive sparse attention and dynamic template updater contribute to superior performance and adaptability in diverse tracking scenarios.