Pneumonia Detection from Chest X-ray Images using Transfer Learning
Shagun Sharma, Kalpna Guleria
Abstract
Pneumonia is a disease that an individual can acquire at any stage of the life. As per a report from Cleveland clinic, America, pneumonia is responsible for about 18% of all contagious diseases. In the phases that follow, this condition may lead to death. Chest radiographs have been found to as constant value by field professionals in clinical practice in order to identify pneumonia. In this work, chest X-ray scans., which are accessible for the identification and diagnosis of pneumonia are utilized. For feature extraction from the images, a VGG16 transfer learning model is used. In this paper, various pre-existing models for pneumonia detection have been reviewed along with the identification of their performance results. A framework of VGG16 used for the study has been explained with a variety of applications of deep learning in disease diagnosis. The dataset for the research is collected from kaggle containing 5., 856 images., which have been further divided into training and testing datasets. Further, the results have been presented in the form of accuracy, precision, Fl-score and recall as 90.8%, 0.9102, and 0.935, 0.9615, respectively.