Automated pathologies detection in retina digital images based on complex continuous wavelet transform phase angles

Abstract

An automated diagnosis system that uses complex continuous wavelet transform (CWT) to process retina digital images and support vector machines (SVMs) for classification purposes is presented. In particular, each retina image is transformed into two one-dimensional signals by concatenating image rows and columns separately. The mathematical norm of phase angles found in each one-dimensional signal at each level of CWT decomposition are relied on to characterise the texture of normal images against abnormal images affected by exudates, drusen and microaneurysms. The leave-one-out cross-validation method was adopted to conduct experiments and the results from the SVM show that the proposed approach gives better results than those obtained by other methods based on the correct classification rate, sensitivity and specificity.

High‐frequency‐based features for low and high retina haemorrhage classification

“…Furthermore, it tolerates high‐dimensional and/or incomplete data [15]. Indeed, the SVM was found to be effective in biomedical classification problems [5, 6, 22, 23].…”

“…Furthermore, it tolerates highdimensional and/or incomplete data [15]. Indeed, the SVM was found to be effective in biomedical classification problems [5,6,22,23]. Though datasets and experiments are different, the results from some related works [1, 7,8] are presented for indication purpose.…”

“…Introduction: Diabetic retinopathy (DR) is a disease that causes blindness; moreover, early and appropriate detection is necessary for better medical treatment. In this regard, several works have proposed automated diagnosis systems of fundus images to assist ophthalmologists in the diagnosis of the disease; either by detecting microaneurysms (MAs) [1][2][3], exudates [4][5][6], haemorrhages (HAs) [1, 7,8], or by vessel segmentation-based approaches [4,9].…”

Particle swarm optimization method for small retinal vessels detection on multiresolution fundus images

“…For example, vessel segmentation is a prerequisite step in the evaluation of retinopathy of prematurity, 6 extraction of measurements on the vessel diameters, 7 detection of the fovea region, 8 detection of arteriolar narrowing, 9 or even in computerassisted laser therapy of retinopathies. 10 Thus, even if many algorithms were proposed in the literature for an automatic detection of retinal lesions in fundus images without retinal vessel segmentation, [11][12][13][14][15] an accurate segmentation of small retinal vessels is still needed for assessment of other diseases.…”

Performance of machine learning methods in diagnosing Parkinson’s disease based on dysphonia measures

“…Machine learning techniques including linear discriminant analysis (LDA) [4], AdaBoost algorithm [5], k nearest-neighbors (k-NN) algorithm and Bayes classifier [6], regression trees (RT) [7], support vector machines (SVM) [8][9][10][11][12][13][14], decision trees (DT), naive Bayes (NB), and multilayer perceptron (MLP) [12] are widely employed in the design of medical decision support systems. Indeed, machine learning techniques have been commonly used in the design of computer-aided-diagnosis (CAD) systems with applications in classification of brain magnetic resonance images [15][16][17], mammograms [18,19], electroencephalography of seizures [20,21], retinal pathologies [22,23], electrocorticogram signals [24], heartbeat signals [25], and arrhythmias [26]. Several machine learning techniques have been employed for supporting the diagnosis of Parkinson's disease (PD) including SVM [1], artificial neural networks (ANN) [27,28], LDA [29], and fuzzy k-NN [30].…”