Automated Classification of Chest Image Using Deep Learning

Abstract

With the current medical research, it has become challenging to diagnose lung and heart diseases. The absolute and verified methodological examination of the patient's medical test's results is necessary for the diagnostic analysis. Deep learning has rapidly enhanced greatly in the past few years, making it capable to recognize and categorize patterns in medical images. This project classifies pneumonia for lung diseases, blocks and contractions in heart for heart diseases. Furthermore, the trained deep learning model is then integrated with a web application that helps the doctors to access and use the model anywhere around the world. The model utilizes convolutional neural networks (CNN) to aid classifications of the diseases based on the patient's clinical data. The experiments conducted and the obtained conclusions and results demonstrate our system may perform efficiently for automated medical picture diagnosis.