Entropy Rate Superpixel Classification for Automatic Red Lesion Detection in Fundus Images

Roberto Romero-Oraá,Jorge Jiménez-García,Javier Oraá-Pérez,María García,María I López-Gálvez,Roberto Hornero

doi:10.3390/e21040417

Abstract

Diabetic retinopathy (DR) is the main cause of blindness in the working-age population in developed countries. Digital color fundus images can be analyzed to detect lesions for large-scale screening. Thereby, automated systems can be helpful in the diagnosis of this disease. The aim of this study was to develop a method to automatically detect red lesions (RLs) in retinal images, including hemorrhages and microaneurysms. These signs are the earliest indicators of DR. Firstly, we performed a novel preprocessing stage to normalize the inter-image and intra-image appearance and enhance the retinal structures. Secondly, the Entropy Rate Superpixel method was used to segment the potential RL candidates. Then, we reduced superpixel candidates by combining inaccurately fragmented regions within structures. Finally, we classified the superpixels using a multilayer perceptron neural network. The used database contained 564 fundus images. The DB was randomly divided into a training set and a test set. Results on the test set were measured using two different criteria. With a pixel-based criterion, we obtained a sensitivity of 81.43% and a positive predictive value of 86.59%. Using an image-based criterion, we reached 84.04% sensitivity, 85.00% specificity and 84.45% accuracy. The algorithm was also evaluated on the DiaretDB1 database. The proposed method could help specialists in the detection of RLs in diabetic patients.

Highlights

Diabetic retinopathy (DR) is a microvascular complication of diabetes leading to vision loss
Following the same approach used in previous studies [32,50], we considered an image pathological when, at least, 30 pixels in the whole image were detected as red lesions (RLs)
The proposed method deals with superpixels instead of pixels to identify the entities of the image

Summary

Introduction

Diabetic retinopathy (DR) is a microvascular complication of diabetes leading to vision loss. As digital color fundus images are low-cost and patient friendly, they are commonly used for large-scale screening [4]. The analysis of these photographs allows the detection of Entropy 2019, 21, 417; doi:10.3390/e21040417 www.mdpi.com/journal/entropy. The strong contrast between the retinal fundus and the black region outside the aperture may yield edge effects in later operations [34]. Pixels outside the aperture were replaced with the mean value of the neighboring pixels inside the FOV [34].

Objectives

Methods

Results

Discussion

Conclusion