| Autor: |
G. Aparna, Vimali J S, Jabez J, Ajitha P, S. Gowri, R. Bharathi |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
2021 5th International Conference on Trends in Electronics and Informatics (ICOEI). |
| Popis: |
Coronavirus disease (COVID-19) is a pandemic caused by the coronavirus SARS -CoV-2 that was not previously seen in humans. COVID-19 is spreading rapidly throughout the world. COVID-19 can be detected by a lung infection of the patients. The standard method for detecting COVID-19 is the Reverse transcription-polymerase chain reaction (RT-PCR) test. But the availability of RT-PCR tests is in short supply. As a result of this, the early detection of the disease is difficult. The easily obtainable modes like X-rays are often used for detecting infections in the lungs. It is confirmed that X-ray scans can be widely used for efficient COVID-19 diagnosis. But a physical diagnosis of X-rays of an outsized number of patients is a longterm process. A deep learning-based diagnosis process can help radiologists in detecting COVID-19 from X-ray scans. Pre-trained CNNs are commonly used in detecting diseases from datasets. This paper proposes a CNN model with a parallelization strategy that extracts the features in the X-ray images by applying filters parallelly through the images. Our proposed method aims to attain higher accuracy and a less loss rate with precision. To do so, the accuracy and loss rates of three types of CNN - VGG-16, MobileNet, and CNN are compared with the parallelization technique. Since, VGG-16 and MobileNet are pre-trained models; those two models are directly imported from Keras. Moreover, this paper utilizes two datasets consisting of COVID X-ray images and Non-COVID X-ray images for the prediction of COVID-19 using Convolution Neural Network [CNN]. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
