v3roi − selects a 3D region of interest(ROI)

v3roi [if=infile] [of=ofile] [ib=] [c=] [s=] [z=] [xl=] [xh=] [yl=] [yh=] [zl=] [zh=] [bb=] [ns=] [-x] [-d] [-D] [-p] [-s] [-x] [-c] [-v]

v3roi selectes a 3D region of interest from a 3D image or 2D sequence of image. The program takes ROI center point location coordinates given by c=(xc[,yc][,zc]) option. If only xc is specified, yc and zc are set equal to xc. The size of the ROI can be specified using multiple options. The first one involves setting individual size parameter (s=(xs[,ys][,zs])) in three dimensions. If only xs is specified, ys and zs are set equal to xs. Individual low and high coordinates in the x,y and z dimensions can be set using xl=,xh=,yl=,yh=,zl=,zh= options.ROI center point z coordinate could be directly set using z= option. As a third option bb= can be used to initialize bounding box parameters. Bounding box parameters can also be extracted from an image possibly a one generated by v3roi using the ib= option. The program extracts the bounding box of the image specified by ib= and set the ROI bbx accordingly.

The size specification options precedence is in the order of extracting bounding box parameters from an input image (ib=) option followed by bounding box specification (bb=), followed by individual parameter initialization (xl=,xh=,yl=,yh=,zl=,zh= or z=), followed by size initialization(xs,ys,zs) considering from higher to lower precedence. The bounding box specification overrides ROI center point coordinate specification (c=) option if both are specified and are conflicting.

Special cases which involve portion of ROI covering region outside the range of the input image are handled using three options. The first option -p keeps the size of the ROI specified by previous options but pads zero in place of the portion of ROI that lies outside the range of the image. The -s option shifts the center of pixel until the ROI is completely inside the input image. The third option (-x) clips the portion of ROI outisde the input image thereby altering the size specification.

Additional options are availbale to use the visionx DICOM information in selecting size of ROI. (-d) option is useful to generate ROI with z dimension is set to be equal to largest dimension in x and y in real space. The resolution of the scan is read from the history header file and is used to calcuate dimensions of the ROI in pixel units. (-D) options copies DICOM header information to the output image if it is available. Cubic ROI in pixel units can be generated using the (-c) option (if -d option is not specified, otherwise it will generate a ROI with equal dimensions in X and Y). The size of the side of the cube is determined from largest dimension specified in x or y direction.

v3roi has also an option that is useful to regenerate a different size ROI from original image given ROI image as an input image (if=). The name of original image is extracted from the history header file of ROI image. The original image should be also available in the same directoy as the input ROI image. Center pixel of ROI is calculated from the input image bbx parameters and a new ROI is extracted from the original image using this center pixel and size parameters specified by ns= (xs[,ys][,zs]) option. This option has a higher precedence over all bbx specification options available.

if=

input file name (accepts any image type).

ib=

input file name to extract bounding box parameters for ROI.

of=

output file name (same as input image type)

c=

(x[,y][,z]) specify center point coordinates of ROI

s=

(xs[,ys][,zs]) size of ROI

bb=

(xl[,xh][,yl][,yh][,zl][,zh]) bounding box specification

xl=

low x coordinate of ROI

xh=

high x coordinate of ROI

yl=

low y coordinate of ROI

yh=

high y coordinate of ROI

zl=

low z coordinate of ROI

zh=

high z coordinate of ROI

z=

ROI center point z coordinate

ns=

(xs,[,ys][,zs]) new size parameters to regenerate ROI

-c

create a cubic ROI in the pixel space (using the largest x or y dim) (affected by anisotropy if -d flag is used)

-s

shift ROI image as much as possible if ROI goes over the image

-x

clip ROI image if ROI goes over the BBX of the input image

-p

pad ROI image with zeros if the ROI goes over the image(default)

-d

if the image is a DICOM derived image, the z dimension number of pixels is calculated using the resolution information from the header to make the ROI isotropic in REAL space.

-D

copy DICOM header in to the output image

A.Asmamaw