Your browser doesn't support javascript.
Forecasting COVID 19 cases on radiological images using deep learning and VGG model
Jundishapur Journal of Microbiology ; 15(1):717-732, 2022.
Article in English | GIM | ID: covidwho-2124772
ABSTRACT
Non covid patients with pneumonia are first analyzed using Chest-X ray (CXR) radiography. But the diagnosis is difficult when analyzing the features of COVID-19 and pneumonia patients since both are having similar features. During this work we have proposed a hypothesis that deep learning can be useful in distinguishing the X ray mages of COVID-19 and pneumonia. can be used as a first-line triage process for non-COVID-19 patients with other forms of pneumonia. The publicly available dataset of COVID 19 is used from Kaggle for evaluating the performance. We have first analyzed the performance using various machine learning algorithms including SVM, KNN, NB, CART etc. Various features used are color and texture descriptors. By using machine learning algorithms we have obtained the accuracy of 81% for 70% training data and 30% evaluation data. Various versions of deep learning models have been used for foresting of COVID 19. Performance is evaluated using One Block VGG, Two block VG, three block VGG, dropout and transfer learning. Performance of One block VGG is 85% with 30 epochs. To summarize the work, we have introduced the use of deep learning techniques for analysis of COVID 19 from chest X ray images. The system is user friendly and rapid.
Keywords
Search on Google
Collection: Databases of international organizations Database: GIM Language: English Journal: Jundishapur Journal of Microbiology Year: 2022 Document Type: Article

Similar

MEDLINE

...
LILACS

LIS

Search on Google
Collection: Databases of international organizations Database: GIM Language: English Journal: Jundishapur Journal of Microbiology Year: 2022 Document Type: Article