Intelligent Automated Detection of Microaneurysms in Fundus Images Using Feature-Set Tuning

Abstract

Due to the widespread of diabetes mellitus and its associated complications, a need for early detection of the leading symptoms in the masses is felt like never before. One of the earliest signs is the presence of microaneurysms (MAs) in the fundus images. This work presents a new technique for automatic detection of MAs in color fundus images. The proposed technique utilizes Genetic Programming (GP) and a set of 28 selected features from the preprocessed fundus images in order to evolve a mathematical expression. Through the binearization of the fitness scores, the optimal expression is evolved generation by generation through a stepwise enhancement process. The best expression is then used as a classifier for real world applications. Experimental results using three publically available datasets validate the usefulness of the proposed technique and its ability to outperform the state of the art contemporary approaches. INDEX TERMS Automatic microaneurysms detection, diabetic retinopathy, fundus images, genetic programming.

Developments in the detection of diabetic retinopathy: a state-of-the-art review of computer-aided diagnosis and machine learning methods

“…On the other hand, in 2020, Usman and Almejalli ( 2020 ) was the first author who was observed innovating an algorithm using a standalone Genetic Algorithm. In that article, they proposed using genetic algorithms (GA) to do General DR Grading.…”

Diabetic Retinopathy Detection and Classification Using Mixed Models for a Disease Grading Database

“…In the context of hybrid methods, various authors have proposed unique hybrid methods for the DR detection process where disparate algorithms are used based on the dataset source, preprocessing, feature extraction, and algorithms for classification methods. [8], Usman and Khalid proposed a new technique for detecting microaneurysms in fundus images using genetic programming (GP), which is also called the intelligent feature set tuning (IFST) technique. They improved feature extraction and their numbers in the preclassification process by generating a mathematical expression for classifying MA images with the MESSIDOR and DIARETDB1 datasets.…”

“…The processing time is a vital parameter in the image retrieval process, and the estimated (this work) processing time is presented in [8] 97.00 98.00 96.00 Zeng et al [10] 95.00 95.00 95.00 Qummar et al [13] 80.80 86.70 63.80 Ramani et al [14] 96.14 92.15 80.31 Ghazal et al [15] 94.00 100 88.00 Soomro et al [30] 95.90 80.20 -Soomro et al [19] 95.60 87.00 -Soomro et al [21] 96.00 75.00 -Soomro et al [20] 96.50 75.00 -Soomro et al [31] 94.70 75.00 -Soomro et al [32] 94.80 73.90 -Soomro et al [22] 94.70 75.00 -Soomro et al [33] 95.20 74.60 96.60 Soomro et al [34] 94.40 --Khan et al [35] 95.01 73.66 96.89 Zago et l. [23] 89.10 --Harangi et al [24] 90.10 --Li et al [25] 92.60 --Proposed work 98.06 83.67 100…”

“…In this process, image processing techniques are generally deployed to extract the retinal fundus image features. Data-mining techniques are applied to build a learning model for classification; this classifier can recognize the presence (or absence) of the disease in retinal fundus images [8]. The implemented CAD methods for DR early detection are familiar and useful and considered a major challenge since they are novel in the medical field.…”

Determinants towards a Better Acceptance Model of IoT in KSA and Eradication of Distrust in Omnipresent Environments

“…In smart homes and buildings, IoT contributes to facilities management, energy management, occupants and resources tracking, and comfort enhancement. On the frontiers of health and automatic disease classification [12], IoT has bolstered through the concepts of ambient-assisted living, wearable devices, internet of mobile things and similar health information systems [13] [14]. In the education sector, IoT has influenced with sensors, intra-communication among wearable technologies, augmented reality and cloud computing for assisted and remote learning [15].…”