| 摘要 |
To provide more correct ribcage boundary information and landmark information, the size of a searching ROI mapped on a digital chest image is decided, the ROI is set on the image, an image existing within the region set on the image is enhanced, a profile of pixel values existing within the region is obtained, and the obtained profile is analyzed so that candidate points for a ribcage boundary is obtained. These deciding, setting, enhancing, obtaining, and analyzing steps are repeated until a center of each searching ROI advances immediately before a given search limit. As a result, a series of candidate points for the ribcage boundary is provided.
|
