Segmentation of macula in retinal images using automated seeding region growing technique

Detection, abstraction and segmentation of the retinal macula is of utmost prominence while studying the effects of macular degeneration. The algorithm deployed uses two steps to pin-point the exact location of the macula. We tested the algorithm on the retinal fundus images acquired from the subsection of MESSIDOR database. Usage of the digital image processing techniques expedite the better extraction of the desired regions. To establish the results, we computed the accuracy, sensitivity and specificity of our outcomes with that of the ground truth images. To segment the macula is our main ambition in high quality fundus images. We provide a novel approach of automated seeding region growing technique which has rarely been used previously. Our algorithm displays great latency in segmenting the large database of affected retinas.