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 8, 2025

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

EDR-Net: Lightweight Deep Neural Network Architecture for Detecting Referable Diabetic Retinopathy.

Ahmad Bukhari Aujih, Mohd Ibrahim Shapiai, Fabrice Meriaudeau

IEEE Transactions on Biomedical Circuits and Systems

|June 14, 2022

Summary

A new deep learning model, EDR-Net, efficiently detects diabetic retinopathy (DR) from fundus images. This computationally lighter architecture achieves performance comparable to state-of-the-art methods, enabling mobile screening applications.

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

Automatic computation of breast cancer biomarkers from multiple [Formula: see text] F-FDG PET image segmentation.

Scientific reports·2026

Same author

Assessing the suitability of automated registration and segmentation for dosimetry calculations in SIRT treatment planning.

EJNMMI physics·2026

Same author

The AROMA Dataset for Automatic Detection of Artifact Type and Severity in Retinal Optical Coherence Tomography Angiography.

Ophthalmic research·2026

Same author

Adaptive noise reduction and speech intelligibility using shape memory metamaterial-based hearing protection earmuffs.

The Journal of the Acoustical Society of America·2026

Same author

Tackling data scarcity: Synthetic tumour and mask generation to improve image segmentation.

Artificial intelligence in medicine·2026

Same author

Extreme cardiac MRI analysis under respiratory motion: Results of the CMRxMotion challenge.

Medical image analysis·2025

Area of Science:

Medical Imaging
Computer Vision
Artificial Intelligence

Background:

Diabetic retinopathy (DR) detection relies on complex convolutional neural networks (CNNs).
Standard CNNs like DR-Net use normal convolution (NC), leading to high computational costs and inefficiency.
Over-parameterization in DR-Net architectures hinders practical application, especially on resource-constrained devices.

Purpose of the Study:

To introduce EDR-Net, an efficient deep learning architecture for diabetic retinopathy detection.
To reduce the computational cost of DR detection models while maintaining high accuracy.
To enable mobile-based computer-assisted screening for referable diabetic retinopathy (rDR).

Main Methods:

Developed EDR-Net, incorporating a depth-wise separable convolution module.

More Related Videos

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

Using Retinal Imaging to Study Dementia

Using Retinal Imaging to Study Dementia

Published on: November 6, 2017

Related Experiment Videos

Last Updated: Sep 8, 2025

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

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

Using Retinal Imaging to Study Dementia

Using Retinal Imaging to Study Dementia

Published on: November 6, 2017

Trained EDR-Net on the DRKaggle-train dataset (35,126 images).

Validated EDR-Net performance on DRKaggle-test (53,576 images) and Messidor-2 (1,748 images) datasets.

Main Results:

EDR-Net achieved predictive performance comparable to state-of-the-art methods for detecting referable diabetic retinopathy (rDR).
The proposed EDR-Net demonstrated at least two times less computation cost compared to other lightweight architectures.
EDR-Net showed superior efficiency, making it suitable for mobile applications.

Conclusions:

EDR-Net offers an efficient and accurate solution for diabetic retinopathy screening.
The reduced computational requirements of EDR-Net facilitate deployment on mobile devices.
This advancement supports accessible and widespread computer-assisted rDR screening.