胸部単純X線写真による間質性肺炎自動診断支援システムの基礎研究

概要

論文の詳細を見る
In this study, we suggest that the method using second-order statistics for distinguishing normal images and interstitial pneumonia images. This method was composed of automatic extraction of lung region, setup of ROIs (Region of interest) to intercostals space, discrimination of ROIs using second-order statistics. The second order statistics used this method is co-occurrence matrix and run-length matrix and the features obtained these matrices can quantify microscopic variation of density. At least, our method was estimated by calculating the ratio of abnormal ROIs. Consequently, we could obtain a relatively good result. From these things, we suggested that this method be useful to the discrimination between interstitial pneumonia images and normal images.
長岡技術科学大学の論文
2003-02-13