Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

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

Dual sagittal-guided coarse-to-fine rib segmentation and numbering in chest CT.

Medical & biological engineering & computing·2026
Same author

A-stream glaucoma shunt versus trabeculectomy: 1-year results on efficacy and safety.

Scientific reports·2026
Same author

A Pilot Study Profiling Aqueous Humor-Derived Exosomal MicroRNA in Primary Open-Angle and Exfoliation Glaucoma.

Ophthalmology science·2026
Same author

Proteomic insights into inflammatory and angiogenic pathways distinguishing uveitic and neovascular glaucoma.

Experimental eye research·2025
Same author

Ultrawidefield-to-Conventional Fundus Image Translation with Scaled Feature Registration and Distorted Vessel Correction.

Bioengineering (Basel, Switzerland)·2025
Same author

Lamina Cribrosa Steepness Index to Measure the Morphology of the Lamina Cribrosa in Myopic Eyes With Optic Disc Distortion.

Translational vision science & technology·2025

Related Experiment Video

Updated: Jan 7, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K

Multi-scale feature enhancement in multi-task learning for medical image analysis.

Phuoc-Nguyen Bui1, Duc-Tai Le2, Junghyun Bum3

  • 1Convergence Research Institute, Sungkyunkwan University, Republic of Korea.

Artificial Intelligence in Medicine
|December 31, 2025
PubMed
Summary
This summary is machine-generated.

This study introduces a novel multi-task learning model for medical image analysis, improving both segmentation and classification accuracy. The ResFormer-based UNet architecture effectively captures local and global features for enhanced disease diagnosis.

Keywords:
Attention mechanismConvolutional neural networksDilated blocksMedical image classificationMedical image segmentationMulti-task learningTransformer

More Related Videos

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.6K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

721

Related Experiment Videos

Last Updated: Jan 7, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K
Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
04:09

Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma

Published on: October 10, 2018

8.6K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

721

Area of Science:

  • Medical Image Analysis
  • Deep Learning
  • Artificial Intelligence

Background:

  • Traditional deep learning models often handle medical image segmentation or classification separately, limiting information sharing.
  • Existing multi-task learning (MTL) methods struggle to balance local context for segmentation and global context for classification.

Purpose of the Study:

  • To develop a unified deep learning model for simultaneous medical image segmentation and classification.
  • To enhance the capture of both local and global contextual information for improved accuracy in both tasks.

Main Methods:

  • A UNet-based multi-task learning model integrating a novel ResFormer block in the encoder for fused local (convolutional) and long-range (Transformer) feature extraction.
  • Multi-scale features from the encoder are combined for classification, while a novel dilated feature enhancement (DFE) module refines decoder skip connections for segmentation.
  • The encoder predicts classification labels, and the decoder generates segmentation masks.

Main Results:

  • The proposed model demonstrated superior performance in both segmentation and classification tasks across multiple medical datasets compared to state-of-the-art methods.
  • The ResFormer block effectively integrates local and global dependencies, enhancing feature representation.
  • The DFE module improved the detection of lesions across various sizes.

Conclusions:

  • The developed UNet-based MTL model with ResFormer and DFE modules offers a significant advancement in medical image analysis.
  • This approach holds potential for improving disease diagnosis and treatment planning through more accurate segmentation and classification.
  • The model's ability to leverage shared information effectively addresses limitations of traditional and existing MTL methods.