Figure 5
From: Automatic image-based segmentation of the heart from CT scans
Position of the spine and the aorta. (a) Common area to all masks computed by thresholding with μsup(k), (b) morphological horizontal dilation of the common area, (c) object of highest area, (d) masked common area (i.e., pixels belonging to the spine), and (e) common area to all masks computed by thresholding with μsup(k) + σ(k) (i.e., pixels belonging to the aorta).