Study of Chest X-rays to Analyse Thorax Diseases using Deep Learning Framework

Abstract

When a patient's doctor suspects they have a lung ailment, one of the first tests they frequently have is a chest X-ray. Sometimes vital information regarding the sickness is missed by radiologists or restorative inspectors. It takes a lot of time to find the sickness' manifestations when these x-beams are reexamined. In this study, we demonstrate how to diagnose thoracic problems using a deep convolutional neural network (CNN). We begin by matching the interest points on the photos to align them. The dataset is then expanded using the theory of Gaussian scale space. The larger dataset is then used to train a deep CNN model in the following step. The larger dataset is then used to train a deep CNN model in the following step. Then, more recent test data that we have are diagnosed using the model. Our research demonstrates how effective our strategy is at producing quality outcomes. Keywords–image classification, machine learning, imageprocessing, neural network.