![]() | Only 14 pages are availabe for public view |
Abstract Strabismus refers to a case in which the two eyes are misaligned. Correct alignment of the two eyes is essential to avoid double vision, to be able of well depth perception and to prevent development or learning delays and consequences of anxiety. The earlier the treatment, the better the results, it might be produced by overactive or underactive eye muscles. Eyeglasses, patching, eye exercises, eye drops, and surgery are the ways for strabismus treatment. An Automated algorithm for Disc Foveal Angle (DFA) calculation and classification of the type of ocular torsion in fundus images is proposed based on digital image processing techniques. For DFA calculation, the Optic Disc (OD) is localized using the Canny edge detector followed by a circular Hough transform. Then, the fovea is localized using the Wellner`s adaptive threshold followed by Canny edge detector and circular Hough transform. The position of the center of fovea with respect to the lower third of OD is used to classify whether the eye is normal or extorsion or intorsion. In our Experiments, a set of 1250 fundus images are used from two publicly available datasets; DRIVE and MESSIDOR and a local dataset from Research Institute of Ophthalmology (RIO). The algorithm is applied on 1250 color fundus images, 40 images acquired from DRIVE database, 1200 images acquired from MESSIDOR database, and 10 images acquired from Research Institute of Ophthalmology. A success rate of 100%, 99.583%, and iv 100% respectively are reached for the optic disc localization an |