IMAGE FEATURE ANALYSIS FOR COMPUTER-AIDED DIAGNOSIS - DETECTION OF RIGHT AND LEFT HEMIDIAPHRAGM EDGES AND DELINEATION OF LUNG FIELD IN CHEST RADIOGRAPHS

Diaphragm edges, together with ribcage edges, in chest radiographs pro vide useful information on the location, shape, and size of the lung f ields that are required by computer-aided diagnosis (CAD) schemes for automated detection of various abnormalities. In this continued study, we developed a computerized method for detection of the right and lef t hemidiaphragm edges. First, the right hemidiaphragm edges in a PA (p ostero-anterior) chest image were determined by edge gradient analysis . An initial vertical ROI was then placed at the middle of the left he midiaphragm, based on a ''standard rule'' for determination of the sta rting points to search for the left hemidiaphragm edges. Seven categor ies were used to assess the accuracy of the placement of the initial R OI and the selection of the primary left starting point within the ini tial ROI. For some categories, it was necessary to select a second lef t starting point besides the primary one. Therefore, for these categor ies, two sets of ''detected left hemidiaphragm edges'' resulted from t he two left starting points. Two parameters were used as measures to e liminate the false left hemidiaphragm edges which were due to an incor rect left starting point. Two polynomial functions were applied separa tely which produced smooth curves for the right and left hemidiaphragm edges. Finally, the delineation of the lung field in a chest image wa s obtained by connecting the right and left hemidiaphragm edge curves with the corresponding ribcage edge curves. The subjective evaluation results indicated that the accuracy for the determination of the right and left hemidiaphragm edges was approximately 97% and 90%, respectiv ely. (C) 1996 American Association of Physicists in Medicine.