Identification of Retinal Image Features using Bitplane Separation and Mathematical Morphology

This paper demonstrates a method for the detection and extraction of features like optic disc and macula from the retinal images. Digital fundus images are becoming popular for the diagnosis of ophthalmic pathologies. Due to this, there is an increasing possibility of applying digital image processing techniques and methods in these images to m the make the diagnosis more easier. In this paper first bit plane separation is done to the pre-processed retinal image. To carry the vital information of the location of optic disc and macula are found by bit plane 0 and bit plane 1. The exact boundary is identified by applying mathematical morphology. The proposed algorithm is simple and has been evaluated on the images collected from a reputed eye hospital. An accuracy of 91% was obtained in extracting the optic disc and macula from the retinal image.