Political Flower pollination Optimizer for age-related macular degeneration detection enabled deep Maxout network using OCT images.

• The ensemble U-Net training is done by PFPO that is formed by merging PO with FPA. • Detection of AMD is carried out by means of DMN, where DMN is trained using PFPO. • The layers are segmented in layer segmentation phase by utilizing ensemble U-Net. Age-related Macular Degeneration (AMD) is the eye state, which influences most elder people. An accurate cause of AMD is yet not understood fully, but it is referred to as multi-factorial with increasing age as the most reliable factor. Thus, AMD prevalence is increasing owing to the aged population in the community. Hence, earlier AMD detection is required to avoid vision loss in the elderly. The detection of AMD in earlier stages at most perils of development permits more appropriate treatment and protection. Even though, arranging a complete eye examination for detection of AMD in the elderly is challenging and laborious. Hence, this research introduces the Political Flower Pollination Optimizer_Deep Maxout network (PFPO_DMN) for AMD detection. Here, median filter is used to pre-process the OCT images. The layers in considered input optical coherence tomography (OCT) images are segmented in the layer segmentation phase utilizing ensemble U-Net. The ensemble U-Net training is accomplished by employing PFPO which is formed by amalgamating the Political Optimizer (PO) with the flower pollination algorithm (FPA). After that, feature extraction is done including texture features, reflectivity, statistical features, thickness and curvature. Lastly, AMD detection is performed using DMN, where DMN is trained by PFPO and hence, detected output is obtained. Furthermore, PFPO_DMN achieved high specificity, sensitivity and accuracy values about 91.6%, 92.9% and 92%. [ABSTRACT FROM AUTHOR]