Severity Quantification of Viral Respiratory Infections in Chest X-Ray Images
Marzieh GolbazOral Defence Date:
Wednesday, May 13, 2015 - 17:00Location:
Committee: Profs. Dragutin Petkovic, Ilmi Yoon and Kaz Okada
Early and accurate assessment of severity of viral respiratory infections (VRIs) is important for improved survival and reduced mortality and morbidity in infants. This study proposes a novel imaging biomarker framework with chest x-ray images to assess the severity of pediatric VRIs. Having the manually segmented lung fields, the large unwanted objects are removed using a graph cut-based segmentation with asymmetry constraint. Each lung field is then subdivided into quadruple areas, allowing the severity of each area to be quantified using information-theoretic heterogeneity measures. The experimental results with a dataset of 148 images and the ground truth severity scores provided by a board-certified pediatric pulmonologist, demonstrates the efficiency and usefulness of the proposed method as a non-invasive imaging biomarker to diagnose and locate lung infections.