Automated Brvo Retinal Detachment Detection Using Artificial Intelligence Techniques

Abstract

Retinal detachment, a very serious eye condition that can result in blindness, occurs when the sensory retina pulls away and separates from the retinal pigment epithelium (RPE). This work focuses on the automatic identification of (Retinal Detachment RD) infundus imaging modality utilized inimage processing techniques. RD is a traumatic condition thatcan lead to blindness. Branch Retinal Vein Occlusion(BRVO) is a smaller retinal venous blockage.  Manually detecting retinal detachment methodsnamely ophthalmoscopic and light spot test are handling very difficult to use.These methods prohibit averted retinal detachment. Identification of retinal detachment the clinical experts to use fundus photographic image to enhance the clinical practice. It is crucial to handling a fundus image with poor quality. The proposed method put to use an integration of level set with region-based active contour methodsFirst, preprocessing steps is applied to retinal imagesfor further processing using Contrast-limited Adaptive histogram equalization (CLAHE) method. Second BRVO segmentation is done by a combined approach of region growing and level set methods. This method usesthe DRIVE and STARE databases for BRVO segmentation.They compared the proposed method with other methods. This method achieves an accuracy of 90%., sensitivity is 1, specificity is 86.9%,precision is 79.3% and f1 score 84%.