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

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

Intravitreal aflibercept 8 mg in patients with neovascular age-related macular degeneration from China: 48-week results from the PULSAR trial.

International journal of retina and vitreous·2026

Same author

Trends of autoimmune diseases in men and inequities in healthcare access and quality in young, working and postworking individuals: a global analysis of 32-year epidemiological study.

BMJ public health·2026

Same author

Arsenic exposure promotes retinal pigment epithelium dysfunction and oxidative stress via NF-κB pathway.

Scientific reports·2026

Same author

Interpretable machine learning to predict functional visual outcomes after the anti-VEGF loading phase for macular edema secondary to retinal vein occlusion: model development and temporal internal validation.

Frontiers in medicine·2026

Same author

The effect of a modified Peyton's four-step teaching method on orthokeratology skills: an experimental study.

Frontiers in medicine·2026

Same author

Association between inflammatory biomarkers and cataracts: validation from the NHANES database and clinical studies.

Frontiers in nutrition·2026

Search research articles

Related Experiment Video

Updated: May 5, 2026

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

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

Published on: November 30, 2022

A vessel-guided multi-task deep learning framework with visual interpretability for simultaneous retinal vessel

Qi Li^1,2, Liming Tao¹

¹Department of Ophthalmology, The Second Affiliated Hospital of Anhui Medical University, Hefei, China.

Frontiers in Medicine

|May 4, 2026

Summary

This study introduces V-MNet, a novel deep learning framework for retinal disease diagnosis. It simultaneously performs vessel segmentation and disease classification, improving accuracy and providing visual transparency for clinical use.

Keywords:

deep learning explainable AI fundus photography multi-task learning ocular disease diagnosis vessel segmentation

More Related Videos

Using Retinal Imaging to Study Dementia

Using Retinal Imaging to Study Dementia

Published on: November 6, 2017

Deep Vascular Imaging in the Eye with Flow-Enhanced Ultrasound

Deep Vascular Imaging in the Eye with Flow-Enhanced Ultrasound

Published on: October 4, 2021

Related Experiment Videos

Last Updated: May 5, 2026

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

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

Published on: November 30, 2022

Using Retinal Imaging to Study Dementia

Using Retinal Imaging to Study Dementia

Published on: November 6, 2017

Deep Vascular Imaging in the Eye with Flow-Enhanced Ultrasound

Deep Vascular Imaging in the Eye with Flow-Enhanced Ultrasound

Published on: October 4, 2021

Area of Science:

Ophthalmology
Medical Imaging
Artificial Intelligence

Background:

Retinal diseases cause significant visual impairment globally.
Accurate diagnosis is vital for preventing vision loss.
Current deep learning models often treat vessel segmentation and disease diagnosis separately, limiting performance.
Lack of transparency in deep learning models hinders clinical adoption.

Purpose of the Study:

To develop a multi-task deep learning framework, V-MNet, for simultaneous retinal vessel segmentation and multi-disease classification.
To enhance model transparency using Grad-CAM for better clinical decision-making.
To leverage the intrinsic relationship between vascular structures and disease pathology.

Main Methods:

V-MNet employs a shared encoder for multi-scale feature extraction.
A segmentation decoder generates retinal vessel masks.
A classification decoder uses a vessel-guided mechanism to transfer structural information for precise pathological region localization.
Integrated Grad-CAM generates class activation maps for interpretability.

Main Results:

V-MNet achieved a Dice coefficient of 0.831 and AUC of 0.985 for vessel segmentation.
The framework obtained an average AUC of 0.978 and F1-score of 0.935 for multi-disease classification.
V-MNet significantly outperformed single-task baseline models and existing state-of-the-art methods.
Ablation studies confirmed the effectiveness of multi-task learning and the vessel-guided mechanism.

Conclusions:

V-MNet offers a promising computer-aided diagnostic tool for retinal diseases.
The vessel-guided multi-task approach effectively utilizes the link between vessel segmentation and disease classification.
Integrated Grad-CAM enhances model transparency, facilitating clinical adoption and decision support.