[Evaluation of an automated adaptive unsharp masking technique in digital chest radiographs]

We are developing a fully automated adaptive unsharp masking technique with parameters that depend on the regional image features of digital chest radiographs. Regions such as the lung fields, retrocardiac area, and spine have extremely different texture patterns and optical densities on chest radiographs. Therefore, it is necessary to enhance the image contrast of each region by an optimum parameter for the regional image. In this study, a chest radiograph was automatically divided into three segments (lung field, retrocardiac area, and spine) by histogram analysis of pixel values. Then, the lung field and retrocardiac area were selectively enhanced with a small mask size and mild weighting factors that had been previously determined as optimum parameters. The spine was enhanced with a large mask size and adequate weighting factors. An observer performance test indicated that this technique provides excellent diagnostic accuracy for simulated nodules over the mediastinum and diaphragm without sacrificing diagnostic accuracy for nodules over the lung field. We believe that this adaptive processing technique may have the potential to be a useful aid for the diagnosis of chest radiographs.