Conjunctivitis Detection using Deep Learning

Conjunctivitis is sometimes called pink eye. The conjunctiva, a thin, transparent tissue that sits over the white portion of an eye, is inflamed. It causes redness, itching, pain, burning, discharge, and swelling in and around the eyes. For effective treatment it is vital to detect the disease accurately as early as possible. This paper composes binary classification to detect whether the eye is normal or affected by the conjunctivitis. The main aim of the paper is, to identify the higher accuracy in detecting the disease. Convolutional Neural Network [CNN] is a variety of machine learning model, and it is also a type of deep learning. This paper uses CNN model such as VGG-16, VGG-19, ResNet-18 and ResNet-152 which is used to detect the conjunctivitis disease. The models are trained by using a large data set that comprises of both normal and affected image of an eye. After the analysing process, it is concluded that the given sample is affected or not. Among the above discussed models VGG-16 has the highest accuracy.