Stage Wise Prediction of Covid-19 Pneumonia from CT images using VGG-16 and SVM

In COVID-19 time, finding medication was the tedious process. Proposed work explains about the segregation of covid-19 CT scan images into categories like mild, moderate and severe on the basis of pneumonia. The dataset uses 227 CT scan images which have been collected manually from hospitals. At fi...

Full description

Saved in:
Bibliographic Details
Main Authors: Bipin Nair, B J, Akash, S, Smaran, R, Hemanth, V, Bhat, Satvik
Format: Conference Proceeding
Language:English
Subjects:
Computational modeling
Computed tomography
COVID-19
Deep learning
Pneumonia
Pulmonary diseases
Support vector machine
Support vector machines
Training
Visual Geometry group model-16
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In COVID-19 time, finding medication was the tedious process. Proposed work explains about the segregation of covid-19 CT scan images into categories like mild, moderate and severe on the basis of pneumonia. The dataset uses 227 CT scan images which have been collected manually from hospitals. At first, the CT scan input images are preprocessed using K-means clustering algorithm. Then Watershed algorithm is used for the segmentation of the pre-processed images to get the affected region. After getting the affected region, VGG-16 model is used for feature extraction process, for model training 53 CT scan images are used as the testing dataset from 185 CT images. Using extracted feature, SVM model will classify the Covid19 pneumonia as mild, moderate, or severe. Finally the classifier has given an accuracy of 96.15% for the prediction of Covid-19 pneumonia stages.
ISSN:2767-7788
DOI:10.1109/ICICT54344.2022.9850529