Stratification of Chest Diseases through Convolution Neural Network

Abstract

For Diagnosis of any problem or condition which is clearly be affecting the lungs or the nearby areas of it, Chest radiographic images are the common used images used by the doctors or radiologist for detecting the type of disease in every medical field. As it is most common and heavily used technology but due to shortage of good and proficient doctors or radiologist the applicability of this application is limited. To rectify this problem a computer based diagnosis by chest radiographic images is designed using Convolution Neural network ( CNN ) by 109,762 front view Radiographic images of 33,816 different patients having multiple scans and each of the patient is diagnose with the any one of the 14 lung diseases which are there in the dataset provided by the National Institute of Health( NIH ) . Hence a CNN model can be developed to classify the disease using the given dataset.