Classification of Infection and Fluid Regions in Chest X-Ray Images
Wan Siti Halimatul Munirah Wan Ahmad, Wan Mimi Diyana Wan Zaki, Mohammad Faizal Ahmad Fauzi, Wooi-Haw Tan · 2016
This paper presents a study on the classification of consolidations in chest radiographs, namely the infection and fluid regions, using a block-based approach with Naïve Bayes classifier. The experiment is performed on infection and fluid regions within the lung, which are divided into 32-by-32 sub-blocks. Several feature extraction techniques are used to capture the block's low level features, and Naïve Bayes classifier is used to categorize each block to either normal, with infection or with fluid. At the region level, the regions are classified into the three categories based on the majority class of the blocks within the region. The performance of the system is evaluated based on its ability to classify the infection and fluid regions into specific abnormalities (infection or fluid or normal) as well as into general abnormality (abnormal or normal). Experimental results show very promising results, with Gabor transform recording the highest overall accuracy.