Related Experiment Video
Updated: Jul 16, 2026
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
EAD-Net: A Novel Lesion Segmentation Method in Diabetic Retinopathy Using Neural Networks
Cheng Wan1, Yingsi Chen1, Han Li1
1Nanjing University of Aeronautics and Astronautics, College of Electronic and Information Engineering, 211106, China.
Insights
Diabetic retinopathy (DR) detection is improved with EAD-Net, a novel convolutional neural network. This automated method accurately segments microaneurysms and hemorrhages, aiding clinical diagnosis and monitoring of this common fundus disease.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Diabetic retinopathy (DR) is a prevalent fundus disease characterized by microvascular lesions like microaneurysms (MAs) and hemorrhages (HEs).
- Manual annotation of these lesions is time-consuming and labor-intensive for clinical analysis.
- Accurate detection and counting of DR lesions are crucial for patient monitoring and diagnosis.
Purpose of the Study:
- To propose a novel segmentation method for accurately detecting and classifying different types of lesions in diabetic retinopathy.
- To develop an automated system that reduces the burden of manual annotation in clinical settings.
- To evaluate the performance of the proposed method against existing techniques and datasets.
Main Methods:
- A convolutional neural network, termed EAD-Net, was developed, comprising encoder, attention, and decoder modules.
- Fundus images underwent normalization and augmentation before automated feature extraction and pixel-wise label prediction using EAD-Net.
- Performance was evaluated using metrics such as sensitivity, specificity, accuracy, and Area Under the Precision-Recall curve (AUPR).
Main Results:
- EAD-Net achieved high performance metrics: 92.77% sensitivity, 99.98% specificity, and 99.97% accuracy on the e_ophtha_EX dataset.
- Comparable AUPR scores were obtained on the IDRiD dataset.
- EAD-Net demonstrated superior performance over the original U-net, particularly in sensitivity and F1-score, with nearly a 10% improvement on a local dataset.
Conclusions:
- The proposed EAD-Net is an effective and novel method for segmenting four distinct types of lesions in diabetic retinopathy.
- The automated segmentation achieved by EAD-Net holds significant clinical importance for the monitoring and diagnosis of DR.
- EAD-Net offers a promising automated solution for lesion detection in DR, outperforming existing methods in key metrics.
Abstract:
Diabetic retinopathy (DR) is a common chronic fundus disease, which has four different kinds of microvessel structure and microvascular lesions: microaneurysms (MAs), hemorrhages (HEs), hard exudates, and soft exudates. Accurate detection and counting of them are a basic but important work. The manual annotation of these lesions is a labor-intensive task in clinical analysis. To solve the problem, we proposed a novel segmentation method for different lesions in DR. Our method is based on a convolutional neural network and can be divided into encoder module, attention module, and decoder module, so we refer it as EAD-Net. After normalization and augmentation, the fundus images were sent to the EAD-Net for automated feature extraction and pixel-wise label prediction. Given the evaluation metrics based on the matching degree between detected candidates and ground truth lesions, our method achieved sensitivity of 92.77%, specificity of 99.98%, and accuracy of 99.97% on the e_ophtha_EX dataset and comparable AUPR (Area under Precision-Recall curve) scores on IDRiD dataset. Moreover, the results on the local dataset also show that our EAD-Net has better performance than original U-net in most metrics, especially in the sensitivity and F1-score, with nearly ten percent improvement. The proposed EAD-Net is a novel method based on clinical DR diagnosis. It has satisfactory results on the segmentation of four different kinds of lesions. These effective segmentations have important clinical significance in the monitoring and diagnosis of DR.
Related Concept Videos
01:27Diabetic Retinopathy
01:22Diabetic Neuropathy