Spotting the Difference Between Healthy Lungs and COVID-19 in Medical Images
Junyoung Kim, Abraham Zepeda, Maria Anjum
Department of Computer Science
Faculty Supervisor: Sara El Alaoui
This study presents a comparative analysis of normal lung images and those of individuals affected by COVID-19 using image classification techniques. Utilizing a convolutional neural network (CNN)-based model, we trained the classifier on a dataset comprising annotated scans of both healthy lungs and COVID-infected lungs. The model was evaluated on its ability to distinguish between the two classes with high accuracy, sensitivity, and specificity. Preprocessing steps such as normalization, data augmentation, and lung region segmentation were applied to improve model performance.