Multi-Class Classification on Chest X-Ray Images Using Convolution Neural Network
18th IEEE India Council International Conference, INDICON 2021
; 2021.
Article
in English
| Scopus | ID: covidwho-1752415
ABSTRACT
The ongoing pandemic caused by the outbreak of the coronavirus has turned human life upside down globally. In this study we present a way of detecting and classifying COVID-19, Viral Pneumonia and Normal classes through chest X-Ray images of the patients using deep learning approach. In this work the database used is Covid-19 Radiography database and we have performed multi-class classification on Chest X-ray images using proposed Customized Convolution Neural Network, Residual Network-50(ResNet-50) and using transfer learning on DenseNet-121 models. Evaluation of the models is done through confusion matrix which further computed the test accuracy, sensitivity/ recall, precision and f1-score. The performance of all the models were quite similar but the Custom CNN performs better with test and validation accuracy of 96% in comparison with the ResNet-50 model which had a validation accuracy of 92% and DenseNet-121 model having a validation accuracy of 80.89% after implementing transfer learning. © 2021 IEEE.
Classification; CNN; Convolution; COVID-19; DenseNet-121; ResNet-50; Transfer learning; Viral Pneumonia; Classifiers; Deep learning; Image classification; X ray radiography; Chest X-ray image; Convolution neural network; Coronaviruses; Multi-class classification; Residual network-50; Classification (of information)
Full text:
Available
Collection:
Databases of international organizations
Database:
Scopus
Language:
English
Journal:
18th IEEE India Council International Conference, INDICON 2021
Year:
2021
Document Type:
Article
Similar
MEDLINE
...
LILACS
LIS