Our paper propose a procedure that effectively detects
affected areas of thoracic region from the X-ray images. Our
deep learning algorithm consisting of Mask-RCNN with
ResNet50 as the backbone localizes the relevant region of
interest with proper augmentation techniques. It can classify
and detect (if any) Pneumothoraxes with an accuracy of
79.29%, which would aid in effective and efficient diagnosis
of the medical condition!