VSTERNVisionX V4 NAME

vstern − sternum segmentation

SYNOPSIS

vstern if= im= of= ca= [-c] [-v] [-d]

DESCRIPTION

vstern segments the sternum from a non-contrast chest CT scan. The inputs are a chest CT and the corresponding label map. The input map is then updated by adding segmented manubrium and sternal body (including the xiphoid process). If -c is specified, the coordinates of the vertical ends and the sternal angle will be output in wish format.

vstern first locates an axial seed slice intersecting the sternal body in an intercostal space. Second, the axial sternum cross section on the seed slice is segmented by matching a rectangle model. Then, it tracks and segments the complete sternum in the vertical direction through sequential axial slices starting from the seed slice. The cross section on each axial slice is segmented using score functions that are designed to have local maxima at the boundaries of the sternum by incorporating intensity and geometry weight. Finally, the sternal angle is localized by analyzing the size profile of the sternum.

The algorithm depends upon the following labels in the image map: skin, airway, rib and vertebra. This version of the algorithm has been evaluated on images with a slice thickness of less than 1.5 mm.

OPTIONS

if=

input non-contrast chest CT scan

im=

input label map (skin, airway, ribs and vertebrae needed)

of=

output updated map (by adding labels of manubrium and sternal body)

ca=

input coordinates of the carina (required format: x,y,z)

-c

optional output of the coordinates of the superior end (sx, sy, sz), inferior end (ix, iy, iz), and the angle (ax, ay, az) of the sternum (in wish format).

-v

verbose mode.

-d

debug mode (save the temporary files).

SEE ALSO

vsribs(1), vsairw(1), vsfas(1), vsvert(1), vsterseed(1), vstertrack(1), vsterrefine(1)

AUTHOR

Shuang Liu, June 2015