Implementation of an AI-Driven Skin Disease Detection System Using MobileNet and Flask Framework

Abstract

One of the most prevalent health problems worldwide is skin disease, and it's important to identify it early for better treatment. However, traditional diagnosis techniques are still largely reliant upon medical dermatologists and extensive manual examination, which isn't always possible for individuals in remote areas or for locations that do not have access to a number of specialists. In this paper, we show how an AI driven skin disease detection system can be built, and yes using deep learning techniques too. The proposed method utilizes mainly the convolutional neural network, MobileNet, which are deployed to facilitate efficient and also relatively accurate classification of different skin diseases from dermoscopic images. The image is preprocessed before the model is run to get improved prediction results and increase the accuracy of the model. Not only that, the web app created is a Flask based web_app and its user-friendly, where a person can upload the skin image and immediately get the results with confidence scores. As a whole, the objective is to minimize human labor, speed up initial screening and facilitate the access to health care through an online platform capable of carrying out automated analysis of skin diseases.