Atlas-based Rib-Bone Detection in Chest X-rays

The National Library of Medicine has developed a portable chest X-ray (CXR) screening system to automatically detect the lung abnormalities in countries where health resources are constrained  [1–3]. The system extracts the texture and shape properties of lung regions from CXR images, and identifies the abnormality using image processing and machine learning algorithms. On a typical CXR, the bone structures overlap with the lung tissue due to the 2D projection of the chest. The rib-cage causes a cross-hatching pattern on the lung region, which misleads the texture analysis  [4–7].

Source: Computerized Medical Imaging and Graphics - April 12, 2016 Category: Radiology Authors: Sema Candemir, Stefan Jaeger, Sameer Antani, Ulas Bagci, Les R. Folio, Ziyue Xu, George Thoma Source Type: research