Chest X-Ray Image Classification using ResNet50 v2

Examining of chest X-rays depends on the experience of radiologists, because the images have no space information and the overlap of different body parts of the chest region may hide infectious tissues. Many of these images are difficult to read when the injury is in low disparity or when they are overlapping with pulmonary vessels. Each chest Xray takes a trained radiologist several minutes to review and give a report, hence increasing the risk of incorrect diagnosis. Current trends shows that Artificial Intelligence is transforming Healthcare from traditional technology to the one in which complex algorithms involving mathematical calculations and formula’s and hardware are used to imitate human perception in the analysis of complicated data. It gives the ability to a radiologist to process an image and gain information in the form of a well-defined output in which the abnormality can be predicted accurately. Image classification using neural networks allows doctors and radiologists to get every point of information from the X-ray image, which is difficult to get when seen through eyes. This paper presents a method to classify images using Deep Residual Networks.