High-Performance Detection of Corneal Ulceration Using Image Classification with Convolutional Neural Networks

Corneal Ulcer, also known as keratitis, represents the most frequently appearing symptom among corneal diseases, second leading cause of ocular morbidity worldwide. Consequences such irreversible eyesight damage or blindness require an innovative approach that enables a distinction to be made between patterns different ulcer stages lower global burden visual disability. This paper describes Convolutional Neural Network-based image classification allows identification types Ulcers based on fluorescein staining images. With balanced accuracy 92.73 percent, our results set benchmark in distinguishing general patterns. Our proposed method is robust against light reflections and automated extraction meaningful features, manifesting strong practical theoretical relevance. By identifying at early stage, we aid reduction aggravation by preventively applying consequently tracking efficacy adapted medical treatment, which contributes IT-based healthcare.