Severity Based Detection of Conjunctivitis and Drug Recommendation System Using CNN

The classification of an eye disease may be useful in the clinical setting for describing the current state of the eye, evaluating the effectiveness of treatment, and selecting the best course of action. Every classification scheme should satisfy the following requirements: clinical character (i.e., be performable by any physician equipped with a standard non-invasive method), repeatability, and meaningfulness (i.e., objective and of clinical relevance to the patient). The image processing approach is appropriate for purely descriptive disease detection. When evaluating the efficacy of a treatment, it is not particularly useful. The easiest way to communicate that result is probably by comparing disease types and stages. To diagnose just one disease, many machine learning algorithms were utilised in the past. Convolutional neural networks (CNNs), for example, Deep learning systems, are capable of inferring a hierarchical depiction of pictures in order to differentiate between proper and improper ocular patterns for diagnostic reasons. In order to assist the doctor (as an aid to doctor) to diagnose viral conjunctivitis at a faster and efficient way based on the severity, we propose our work to use Multi-dimensional (Md-CNN) architecture of conv-net for classification of medical eye image processing which has a higher efficiency in-case of medical image processing and to provide suggestions on medication based on the severity of the infection.