Pleural Effusion Classification From Chest X-Ray Images Using Sift And Zernike Features

Abstract

Pleural effusion is an uncommon lung disease characterized by a build-up of fluid between the two layers of the pleura that causes specific symptoms, such as chest pain and shortness of breath. Machine learning techniques have been widely used for abnormality detection in medical images. Chest X-ray images (CXR) are among the non-invasive diagnostic tools used to detect various disease. In the proposed work, two different feature extractions namely Scale Invariant Feature Transform and Zernike features are extracted. The extracted features are fed into the classifier namely Random Forest and K-nearest Neighbor which classifies into effusion, emphysema, infiltration, no-findings and pleural thickening. The results are compared were ZERNIKE with Random Forest grants the satisfactory results of 94.30 %.