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: Sep 2, 2025

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Attention-augmented U-Net (AA-U-Net) for semantic segmentation.

Kumar T Rajamani¹, Priya Rani², Hanna Siebert³

¹Philips Research, Bangalore, India.

Signal, Image and Video Processing

|August 1, 2022

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

Beyond the LUMIR challenge: The pathway to foundational registration models.

Medical image analysis·2026

Same author

Underutilisation of hydroxyurea in sickle cell disease: a global scoping review of multilevel barriers and facilitators.

Journal of community genetics·2026

Same author

CoRe: Joint Optimization with Contrastive Learning for Medical Image Registration.

Sensors (Basel, Switzerland)·2026

Same author

Evaluation of Intercondylar Distance as a Predictor of Vertical Dimension in Edentulous Patients: A Cross-sectional Comparative Study.

The journal of contemporary dental practice·2026

Same author

HuR enhances SARS-CoV-2 non-structural protein translation through the genomic 5'-UTR, by promoting polypyrimidine tract-binding protein binding.

Journal of virology·2026

Same author

Prevalence and Predictors of Rapid Estimated Glomerular Filtration Rate Decline in Patients with Non-albuminuric Diabetic Kidney Disease: A Single-centre Retrospective Cohort Study.

The Nigerian postgraduate medical journal·2026

Same journal

Accumulated bispectral image-based respiratory sound signal classification using deep learning.

Signal, image and video processing·2023

Same journal

Control learning rate for autism facial detection via deep transfer learning.

Signal, image and video processing·2023

Same journal

SAM C-GAN: a method for removal of face masks from masked faces.

Signal, image and video processing·2023

Same journal

N-Net: an UNet architecture with dual encoder for medical image segmentation.

Signal, image and video processing·2023

Same journal

A novel and accurate deep learning-based Covid-19 diagnostic model for heart patients.

Signal, image and video processing·2023

Same journal

Machine learning approach for detecting Covid-19 from speech signal using Mel frequency magnitude coefficient.

Signal, image and video processing·2023

See all related articles

A novel attention-augmented U-Net (AA-U-Net) improves COVID-19 CT image segmentation by better capturing spatial context. This deep learning model enhances lesion detection accuracy with limited data.

Area of Science:

Medical Imaging
Artificial Intelligence
Computer Vision

Background:

Deep learning models require extensive spatial context for accurate image segmentation.
Limited labeled data for COVID-19 CT segmentation poses a significant challenge for existing models.
Attention mechanisms, particularly self-attention, enhance contextual information capture in deep networks.

Purpose of the Study:

To develop a novel attention-augmented convolution U-Net (AA-U-Net) for improved COVID-19 lesion segmentation on CT images.
To enhance the spatial aggregation of contextual information in encoder-decoder segmentation architectures.
To address the scarcity of labeled data in COVID-19 medical image analysis.

Main Methods:

Integration of attention-augmented convolution modules within the bottleneck of a U-Net architecture.

Keywords:

Attention mechanism Attention-augmented convolution COVID-19 Consolidation Ground-glass opacities Segmentation U-Net

More Related Videos

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Sep 2, 2025

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Utilizing concatenated self-attention and convolution feature maps to capture long-range interactions.

Training and validation of the AA-U-Net model on COVID-19 CT datasets for lesion segmentation.

Main Results:

The AA-U-Net achieved Dice scores of 72.3% for ground-glass opacity and 61.4% for consolidation lesions.
Demonstrated a performance improvement of 4.2% over a baseline U-Net and 3.09% over a U-Net with matched parameters.
The model's enhanced performance is attributed to its ability to capture dynamic and precise attention context.

Conclusions:

The proposed AA-U-Net significantly improves semantic segmentation accuracy for COVID-19 lesions in CT images.
Attention-augmented convolution effectively enhances contextual information aggregation, crucial for segmentation tasks with limited data.
This approach offers a promising solution for medical image segmentation in resource-constrained scenarios.