ABSTRACT
Respiratory diseases have been known to be a main cause of death worldwide. Pneumonia and Covid-19 are two of the dominant diseases. Several deep learning based studies are available in the literature that classifies infection conditions in chest X-ray images. In addition, image segmentation has been also applied to obtain promising results in deep learning approaches. This paper focuses on using a modified version of the U-Net architecture to conduct segmentation on chest X-rays and then use segmented images for classification to assess the impact on the performance. We achieved an Intersection over Union of 93.53% with the proposed modified U-Net architecture and achieved 99.36% accuracy on segmentation aided ensemble classification.
ABSTRACT
Respiratory diseases have been a main reason for death in many countries worldwide. This study considers Pneumonia which is a common lung infection condition and COVID-19 which was declared a pandemic in 2020. Since both diseases can lead to life-threatening conditions, detecting these conditions at an early stage is crucial to properly treat the patients. While chest X-rays are widely used for diagnosing these diseases, it requires expert knowledge. This study focuses on introducing a deep learning based approach for analysing chest X-ray images to detect normal, Pneumonia and COVID-19 conditions. Experiments were conducted with multi-model deep learning models including MobileNetV2, Resnet50, InceptionV3, and Xception architectures with added layers, and 5-fold cross-validation. The results of ResNet50 show an average accuracy and recall of 98.87% and 98.54%, respectively. © 2022 IEEE.
ABSTRACT
Diseases in the respiratory system affect many people worldwide and can lead to life-threatening conditions. Pneumonia is an acute infection of the lungs and Coronavirus is a recently emerged respiratory disease that has been recorded in many deaths around the world and announced as a pandemic in early 2020. It is crucial to detect these conditions at an early stage as possible for proper treatment. Among many treatment strategies, chest X-rays are widely used in the diagnostic process. This study presents a deep learning based approach to analyse chest X-ray images to distinguish normal Pneumonia or COVID-19 Pneumonia conditions. We have followed the MobileNetV2 architecture with additional layers added to the top of the architecture. Our results show an average accuracy of 98.65% and an average recall of 98.15% with 5-fold cross-validation.