A Rule-Based Method for Detection of Ribcage Boundary in Chest Radiographs(International Forum on Medical Imaging in Asia 2009 (IFMIA 2009))

概要

論文の詳細を見る
The ribcage boundary of chest images provides useful information on the location, shapes, and size of lung fields that are required by computed-aided diagnosis schemes for lung nodule detection. But, the previously reported methods for detection of ribcage boundary often fail to detect the top lung part of the ribcage boundary. In this paper we propose a novel method for detection of ribcage boundary in chest images. In experiments using 247 images of JSRT database which is a public database, the proposed method could detect the complete ribcage boundary including the top lung part in good accuracy for 241 images and in fair accuracy for 4 images. The time needed by the proposed method to detect the entire ribcage boundary for each image of 512x512 matrix size was about 0.25 seconds on the average by a personal computer with 1.4GHz CPU.
社団法人電子情報通信学会の論文
2009-01-12