Image feature analysis and computer-aided diagnosis in digital radiography: automated detection of pneumothorax in chest images.

In order to aid radiologists in the diagnosis of pneumothorax from chest radiographs, an automated method for detection of subtle pneumothorax is being developed. The computerized method is based on the detection of a fine curved-line pattern, which is a unique feature of radiographic findings of pneumothorax. Initially, regions of interest (ROIs) are determined in each upper lung area, where subtle pneumothoraces commonly appear. The pneumothorax pattern is enhanced by the selection of edge gradients within a limited range of orientations. Rib edges included in this edge-enhanced image are removed, based on the locations of posterior ribs that are determined separately. A subtle curved line due to pneumothorax is then detected by means of the Hough transform. The detected pneumothorax pattern is marked on the chest image displayed on a CRT monitor. With the present computer method applied to 50 chest images (28 normals and 22 abnormals with pneumothorax), we were able to detect 77% of pneumothoraces, with 0.44 false-positives per image.