CataractEyeNet: A Novel Deep Learning Approach to Detect Eye Cataract Disorder

Humans see the happenings around through the help of eyes. Currently, visual impairment and blindness have become significantly dangerous health problems. Even though advanced technologies are emerging rapidly, blindness and visual impairment, still, remain significant problems around the globe in healthcare systems. Specifically, cataract is among the problem that results in poor vision and may also cause falling as well as depression. In old times, mostly the old people were suffering; however, childhood cataracts are common that result in severe blindness as well as visual impairment in children as well. Therefore, it is extremely mandatory to develop an automated system for the detection of cataracts. Being that this research presents a novel deep learning-based approach, CataractEyeNet to detect cataract disorder using the lens images. More specifically, we customized the pre-trained VGG-19 model and added 20 more layers to enhance the detection performance. The CataractEyeNet has obtained an accuracy of 96.78%, precision, recall, and F1-score of 97%, 97%, and 97%, respectively. The experimental outcomes of the CataractEyeNet show that our system has the capability to accurately detect cataract disorders.