Honey Badger Aquila optimization-based deep learning with multi-kernel shape index histograms for diabetic macular edema classification

Abstract

Diabetic Macular Edema (DME) is the foremost reason for vision loss in patients having Diabetic Retinopathy (DR). In the earlier diagnosis of DR, Optical Coherence Tomography (OCT) plays a major part in detecting and classifying DR, thus preventing vision loss. Most people suffer from DME due to neglecting treatments, which may lead to blindness or visual impairment. If properly detected, this can be healed at an earlier phase. DME detection and classification is a challenging chore in affected patients. To vanquish the. Challenging task, an effective method is proposed for DME detection as well as classification using the proposed Shape Index Histogram Honey Badger Aquila Optimization-based deep convolutional neural network (SIH+HBAO-based deep CNN). Pre-processing is conducted employing a Gaussian filter. After pre-processing, layer segmentation is conducted by Correlative-based gradient global thresholding with active contour. Then, feature extraction is performed whereas layer-specific features, texture features including Local Gradient Pattern (LGP) and proposed SIH with multi-kernel are extracted. Furthermore, the proposed SIH with multi-kernel feature is devised by modification of shape index histogram with multi-kernel. After that, DME detection and classification are conducted utilizing Deep CNN, which is tuned employing the proposed HBAO algorithm. The proposed HBAO algorithm is introduced by incorporation of Honey Badger Algorithm (HBA) and Aquila Optimizer (AO). Moreover, the proposed SIH+HBAO-based deep CNN attained maximal values of accuracy, sensitivity and specificity of 0.912, 0.913 and 0.917.