Algoritmo para segmentação de drusas em imagens de retinografia através de técnicas de processamento digital de imagens

Abstract

The increase in life expectancy in recent decades has created significant new challenges to public health worldwide, which are associated with diseases related to the aging of the population. Among these conditions, Age-Related Macular Degeneration (AMD) represents one of the main causes of progressive vision loss and irreversible blindness in the elderly. AMD is diagnosed through ophthalmological examinations, such as retinography, which consists of capturing high-resolution images of the back of the eye. This examination allows the visualization of the structures of the retina and the identification of changes, such as drusen, which represent one of the main markers of the disease. Early detection of AMD enables interventions that slow the progression of total vision loss. Thus, this work developed an algorithm with the objective of assisting professionals in visualizing the region and identifying drusen, based on the segmentation of drusen in retinography images, using basic digital image processing techniques. Basic digital image processing techniques were applied to develop the strategy, such as: thresholding, mathematical morphology, histogram equalization and k-means segmentation. After extracting the region of interest, the results were compared with other studies published in the literature in order to evaluate the effectiveness and performance of the proposed method. The results achieved, with an accuracy of 97.53% and specificity of 99.18%, indicated a high degree of accuracy and reliability in the performance of the algorithm.