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 26, 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

DenseViT-OCT: A Hybrid CNN-Transformer Architecture with Multi-Scale Dense Feature Aggregation for Automated

Elif Yusufoğlu¹, Salih Taha Alperen Özçelik², Orhan Atila³

¹Department of Ophthalmology, Elazig Fethi Sekin City Hospital, 23100 Elazig, Turkey.

Tomography (Ann Arbor, Mich.)

|June 25, 2026

Summary

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

Deep Learning-Assisted Three-Dimensional Segmentation of Vertebrobasilar Artery Calcification in Cone Beam Computed Tomography.

Journal of imaging informatics in medicine·2026

Same author

Query-Driven Retinal Layer Segmentation in OCT Using Cross-Attentive Feature Learning.

Diagnostics (Basel, Switzerland)·2026

Same author

Automatic Infant Movement Assessment Using Pose-LBP Features and a Cost-Sensitive Subspace kNN Ensemble.

Bioengineering (Basel, Switzerland)·2026

Same author

An Explainable Plane-Wise ConvNet Approach for Detecting Femoral Head Osteonecrosis from Magnetic Resonance Images.

Bioengineering (Basel, Switzerland)·2026

Same author

Explainable Neutrosophic Knowledge Distillation Model for Ocular Disease Classification Using Ultra-Wide Field Fundus Images.

Bioengineering (Basel, Switzerland)·2026

Same author

Grad-CAM Enhanced Explainable Deep Learning for Multi-Class Lung Cancer Classification Using DE-SAMNet Model.

Diagnostics (Basel, Switzerland)·2026

This summary is machine-generated.

A new deep learning model, DenseViT-OCT, automates epiretinal membrane (ERM) severity classification from OCT scans with high accuracy. This method offers a reliable alternative to manual grading for improved patient care.

Area of Science:

Ophthalmology
Medical Imaging
Artificial Intelligence

Background:

Epiretinal membrane (ERM) is a common vitreoretinal disorder causing visual impairment.
Accurate ERM severity grading from optical coherence tomography (OCT) is crucial but challenging due to manual subjectivity.
Automated classification methods are needed to improve efficiency and consistency.

Purpose of the Study:

To develop and validate a deep learning-based automated method for classifying ERM severity.
To enhance the accuracy and reliability of ERM grading using OCT images.
To provide a decision-support tool for clinical management of ERM.

Main Methods:

A hybrid deep learning model, DenseViT-OCT, integrating CNNs and ViTs was developed.

Keywords:

attention mechanism computer-aided diagnosis deep learning epiretinal membrane hybrid CNN-Transformer medical image classification multi-scale feature fusion optical coherence tomography retinal imaging vision transformer

Related Experiment Videos

Last Updated: Jun 26, 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

Key modules include Multi-Scale Dense Feature Aggregation (MDFA), Adaptive Feature Calibration (AFC), and Cross-Attention Feature Fusion (CAFF).

The model was trained and validated on 2195 OCT B-scan images from 397 patients and externally validated on the OCTDL dataset.

Main Results:

DenseViT-OCT achieved 94.76% accuracy on an internal test set, outperforming 19 benchmark models.
The model demonstrated strong performance with macro-averaged precision (93.76%), recall (93.22%), F1-score (93.47%), and AUC (98.95%).
External validation showed 90.76% accuracy and 97.61% AUC, indicating good generalization.

Conclusions:

DenseViT-OCT offers a robust framework for automated ERM severity classification from OCT B-scans.
The hybrid approach combining CNN and ViT features improves classification performance.
The method shows promise as a research-oriented decision-support tool for ERM assessment, warranting further clinical validation.