Cost Sensitive Learning using Chest X-Ray with CNN for Covid -19 Detection with Lung Diseases which Lead to Class Imbalance

Due to problems with class balance, utilizing a chest X-ray to find Covid-19 has become a very difficult task. This research suggests utilizing Convolutional Neural Networks (CNNs) to resolve the issue with class imbalance and increase the performance of Covid-19 identification. The suggested solution employs a CNN model made especially for the issue and cost-sensitive learning to solve it. In this investigation, chest X-ray pictures of people with Covid-19 and lung illnesses were employed. The dataset was quite uneven, and most of the photos were normal. There were a few COVID-19 and pneumonia images, however. In this research, cost-sensitive learning is used, which contains three distinct ways (i) data resampling, (ii) algorithm changes, and (iii) ensemble methods, to fix the uneven distribution and prevent getting biased classification results. Experimental findings display that the suggested technique outperforms previous state-of-the-art methods, with a greater rate of detecting Covid-19 and superior performance measures including accuracy, recall, and F1-score. With an accuracy of $91.67 \%$ and a sensitivity of $88.89 \%$, the suggested method shows promise. The feature maps produced by the CNN model to figure out which parts of the chest X-ray pictures help to detect Covid-19. It has been suggested that cost-sensitive learning is an ideal way to deal with the class imbalance issue in Covid-19 detection using medical imaging. In the research report, the suggested method worked well for finding Covid-19 in chest X-rays of people with lung disorders that cause class imbalance. However, further research is required to confirm its generalizability and therapeutic usefulness.