Related Experiment Video
Updated: Oct 20, 2025
07:22
Quantitative Fundus Autofluorescence for the Evaluation of Retinal Diseases
Published on: March 11, 2016
11.6K
Red-lesion extraction in retinal fundus images by directional intensity changes' analysis
Maryam Monemian1, Hossein Rabbani2
1School of Advanced Technologies in Medicine, Medical Image & Signal Processing Research Center, Isfahan University of Medical Sciences, 8174673461, Isfahan, Iran.
Scientific Reports
|September 15, 2021
Summary
A new method accurately detects red lesions in the retina, crucial for diagnosing diabetic retinopathy (DR). This technique offers high sensitivity and specificity across multiple datasets, aiding in disease severity assessment.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer Vision
Background:
- Diabetic retinopathy (DR) is a significant complication of diabetes, characterized by retinal blood leakage and the formation of red lesions.
- Analysis of these red lesions is critical for determining the severity of diabetic retinopathy.
Purpose of the Study:
- To propose a novel and efficient method for extracting red lesions from retinal fundus images.
- To accurately identify and discriminate red lesions from other features in the retina.
Main Methods:
- The proposed method identifies boundary pixels of blood vessels and red lesions.
- It utilizes distinguishing features of red lesion boundary pixels, leveraging significant intensity changes in fundus images.
- Special neighborhood windows are employed around extracted boundary pixels to enhance detection.
Main Results:
- The method achieved high sensitivity and specificity across three datasets: Diaretdb0 (0.89, 0.9), Diaretdb1 (0.87, 0.88), and Kaggle (0.82, 0.9).
- The red-lesion extraction process demonstrated time-efficient performance.
Conclusions:
- The novel red-lesion extraction method is effective and efficient for analyzing diabetic retinopathy.
- This technique shows potential for improving the diagnostic process and severity assessment of DR.
