Kvasir – Automated Gastro-Intestinal Disease Classification.

Abstract

GI diseases are a major health concern in the world today and calling for prompt and accurate diagnosis for treatment. This study intends to capture the increasing incidence of GI diseases by devising an early detection and management mobile application using deep learning. It combines medical image analysis with patients’ data to forecast GI conditions with relative accuracy, via features of image data. After testing the model for testing and evaluating different models of Deep learning, the best one is chosen for the diagnosis of GI diseases. The mobile application is intended to be very easy to use with an account creation page, a login page and a diagnosis page where uploaded medical images are sorted, classified, and predictions about potential GI diseases are made. Flask is used to develop the API to interconnect the mobile app and the deep learning model so that the real time predictions can be made as well as recommendations. The integration will help healthcare workers and clients get early warning signs to seek doctor’s checkup in case of possible GI complications. By linking between the theoretical models and the application, this project aims at making a positive impact on the direction and efficient care of managing GI diseases and the patient’s experience through it.