Detection of Coronavirus Disease (COVID-19) Through X-Ray Images Using K-Nearest Neighbor Classifier Based on DBSCAN
摘要
The global COVID-19 pandemic has posed a significant public health challenge, with over 479 million reported cases worldwide as of March 2022. This crisis has led to a scarcity of commercial testing equipment, emphasizing the need for an automated system to detect this respiratory infection, particularly in hospitals facing an escalating daily caseload. To address this, our study proposes a K-Nearest Neighbor (KNN) classifier based on DBSCAN for detecting COVID-19 pneumonia in patients using chest X-ray images. The results are binary, indicating either a Positive or Negative diagnosis for COVID-19. The algorithm assesses the test images against the training images using the K-neighbors’ distances. Notably, our findings underscore how variations in K-neighbors influence classification performance, achieving a precision below 70%. A critical aspect of program development involves accurate labeling of training set images, as this significantly impacts the model’s ability to classify medical images.