Quantitative diagnosis of pneumoconiosis based on recognition of small rounded opacities in chest X-ray images

A method is presented for automated categorization of pneumoconiosis chest radiographs. The method is based on recognition of individual lesions in a film. The processing procedure consists of the following four steps: extraction of candidate regions of small rounded opacities, recognition of abnormal opacities, calculation of the opacity density and film categorization. The opacity density is defined by the area of the number of the opacities detected within a certain lung region. In the experiments using ILO international standard films, the opacities detected by computer were highly coincident with those traced by an experienced radiologist, and the density values calculated were also strongly correlated with the classification code of the standard films