dc.description.abstract |
Corona virus disease (COVID-19), which is caused by the severe acute respiratory
syndrome corona virus 2 (SARS-CoV-2), is an inflammatory disease that causes
respiratory illness (similar to influenza) with symptoms such as cold, cough, and fever, as
well as difficulty breathing in more severe cases. The goal of the research is to develop a
model that can identify COVID-19. The information is gathered from hospitals. The visual
data was analyzed using a deep learning approach (CNN).
Preprocessing, feature extraction, and classification are the three components of the
research. We normalize the image to a standard size during image processing. For feature
extraction, we employ a convolutional neural network. It is used to identify and choose
essential elements that contribute to the disease's symptom. We employ a convolutional
neural network for classification. For classifying into a specific class (normal/no findings,
pneumonia, and COVID-19), a three-way Softmax is employed. The research was carried
out in Python using Keras (with TensorFlow as a backend) and evaluated on a sample
image dataset obtained from Tibebe Gion Hospital.
The model achieved a diagnosis accuracy of 98.35% for training and 98% for testing to
identify Covid-19. This research work presented different contributions that can be further
improved or implemented on the effort to detect and grade related diseases.
Keywords: Covid-19, Deep Learning, CNN, Feature Learning, SoftMax |
en_US |