Automated Localization of Centre of Optic Disc and Centre of Macula in Retinal Images

An automated system is presented to locate the centre of an optic disc (OD) and the centre of macula in the retinal image This work is carried out for control point detection required for registration of retinal images. Control point detection is an important step for registration. Unique feature points in the image are used as control points. Optic disc and macula are unique anatomic landmarks within retinal image. Hence, centre of optic disc and centre of macula can be used as control points for registration. Registration is an important step in super-resolution of retinal images and to identify change in the images due to presence of diseases. The propose methods can handle optic disc and macula in images of both the left and right eye. The system first detects optic disc and macula in retinal image. This paper proposes an efficient combination of algorithms which avoids false detection of optic disc and macula. An ensemble of different algorithms based on different principles produce  ore accurate results. The optic disc is the brightest circular region whereas macula is the darkest circular region within the retinal image. These features are used for detection of optic disc and macula. The system is developed and tested on a set of 60 images. It is able to find the optic disc location and its centre in 100% and the macula and its centre in 100% of all tested cases.