Related Experiment Video
Updated: Nov 27, 2025
07:15
Transient Optical Clearing Using Absorbing Molecules for Ex Vivo and In Vivo Imaging
Published on: July 11, 2025
1.5K
Entropy Rate Superpixel Classification for Automatic Red Lesion Detection in Fundus Images
Roberto Romero-Oraá1, Jorge Jiménez-García1, María García1
1Biomedical Engineering Group, E.T.S.I. de Telecomunicación, University of Valladolid, 47011 Valladolid, Spain.
Entropy (Basel, Switzerland)
|December 3, 2020
Summary
This study presents an automated method for detecting red lesions in retinal images, crucial for early diabetic retinopathy diagnosis. The system achieved high accuracy, aiding specialists in identifying early signs of the disease.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Diabetic retinopathy (DR) is a leading cause of blindness in working-age adults.
- Early detection of DR through identifying red lesions (RLs) like hemorrhages and microaneurysms is critical.
- Automated systems can enhance large-scale screening for DR.
Purpose of the Study:
- To develop an automated method for detecting red lesions (RLs) in digital color fundus images.
- To aid in the early diagnosis and screening of diabetic retinopathy.
- To improve the efficiency and accuracy of DR lesion detection.
Main Methods:
- A novel preprocessing stage was employed for image normalization and retinal structure enhancement.
- The Entropy Rate Superpixel method was utilized for segmenting potential RL candidates.
- A multilayer perceptron neural network was used for classifying superpixels.
Main Results:
- The developed algorithm achieved 81.43% sensitivity and 86.59% positive predictive value on a pixel-based criterion.
- An image-based criterion yielded 84.04% sensitivity, 85.00% specificity, and 84.45% accuracy.
- The method demonstrated effectiveness when evaluated on the DiaretDB1 database.
Conclusions:
- The proposed automated method shows promise for detecting red lesions in diabetic retinopathy.
- This system can assist specialists in the early identification of DR signs in patients.
- Automated analysis of fundus images offers a valuable tool for DR screening and diagnosis.
