Automated Localisation of Optic Disc and Macula from Fundus Images

Optic disc (OD) detection is a main step while developing automated screening systems for diabetic retinopathy. Optic disc boundary and localization of macula are the two features of retina was necessary for the detection of exudates and also knowing the severity of the diabetic maculopathy. According to the prior information, the diameter of optic disc in a standard retinal image, connected components and iterative thresholding was used to locate optic disc. The macula was localized based on its distance and position with respect to the optic disc as it remained relatively constant. Even though macula is considered to be one of the darkest regions without vessels in a retinal image, less contrast between the macula and background makes it difficult to locate based on image variance. Among 148 images considered for evaluating the methods optic disc and macula were localized with sensitivity of 99.32% and 96.6% respectively. Detection of optic disc boundary becomes important for the diagnosis of glaucoma. Geometric active contour model was explored to segment the optic disc boundary as segmentation algorithms failed to provide good result. Image segmentation was performed by starting with initial curve and evolving its shape by minimizing energy function represented by level set function. The iterative curve evolution was stopped at the image boundaries where the energy was minimum. Experiment was performed on both RGB image and gray scale image and found that implicit active contours provided better result with gray scale images. Total of 74 images were used to evaluate the method. Optic disc boundary drawn manually by an expert was used as ground truth. The method was able to achieve average sensitivity of 90.67% with mean of ±5.05. Based on the result obtained in optic disc boundary detection, it can be stated that geometric based implicit active contour models provide a better segmentation for images with weak boundaries when compared to parametric models. Keywords— Optic disc detection, fovea, image processing, fundus image, diabetic retinopathy