Abstract: The year 2020 has witnessed the effects of global pandemic outbreak through the unprecedented spread of novel coronavirus COVID-19. As the testing of coronavirus happened manually in the initial stage, the ever- increasing number of COVID-19 cannot be handled efficiently. Also, the coronavirus is divided into 3 phases and it has different effectson lungs. To handle this situation, researchers have attempted to detect coronavirus using chest X-ray images and Chest CTscan images by using Artificial Intelligence[AI] technologies. AI helps to forecast the coronavirus cases for analysing the virus structure and chest X-Ray and CT scan images helps to predict the stags of corona virus. Henceforth, this paper has developed a CNN model, which utilizes 3 classes as follows: positive COVID-19 images, normal images and viral pneumonia images. The model has been trained on these set of images and got 94% of accuracy on training dataset and 96% of accuracy on validation dataset. The proposed model has achieved the test accuracy of 94% for 3 classes in Chest X-Ray image classification. The main motive behind developing this model is to reduce its computational time by using less layers and more hyper parameter tuning. The proposed model is compared with pre-existing models as they were more complexand took much training time. Till now 94% of accuracy has been achieved on test dataset.

Keywords- Convolutional neural networks, COVID-19, neural networks, X-Ray

PDF | DOI: 10.17148/IJARCCE.2023.125205

Open chat
Chat with IJARCCE