Proposal of a Nodule Density-Enhancing Filter for Plain Chest Radiographs on the Basis of the Thoracic Wall Outline Detected by Hough Transformation(Special Issue on Measurements and Visualization Technology of Biological Information)

概要

論文の詳細を見る
In screening for primary lung cancer with plain chest radiography, computer-aided diagnosis systems are being developed to reduce chest radiologists' task and the risk of missing positive cases. We evaluated a difference filter that enhances nodule densities in the preprocessing of chest X-ray images. Since ribs often affect detection of pulmonary nodules, we designed an eye-shaped filter to fit the rib shape. Although this filter increased the nodule detection rate, it could not detect nodules near the thoracic wall. The thoracic wall was then outlined by computers with Hough transformation for line detection. On the basis of the outline, the direction of the eye-shaped filter was determined. With this technique, the filter was not affected by considerable changes in the shape of anatomical structures, such as ribs and the thoracic wall, and could detect pulmonary nodules regardless of their location.
社団法人電子情報通信学会の論文
2002-01-01