| Author: |
J. Anthony Parker, MD PhD (J.A.Parker@IEEE.org ) |
| Dates: |
2003/04/07: First version |
| Source: |
The source file, Projector_TP.java. |
| Installation: |
Download Projector_TP.class
to the plugins or "plugins/NucMed" folder and restart ImageJ. |
| Description: |
This ImageJ PlugInFilter is a minor modification of the
ImageJ class Projector (Image/Stacks/3D Project...). Five
features have been added:
1. Depth cueing has been added to the mean value option.
The cueing decreases the value of a pixel as a negative
exponential function of depth. Reasonable values for this
parameter are from 0 to about 1 or so.
2. "Z-Axis Orthogonal" has been added as a type of Axis of
Rotation. The normals to the projections for the "Z-Axis"
option are parallel to the z axis. The normals to the projections
for the "Z-Axis Orthogonal" option are all perpendicular to the z axis.
In Radiology, given an axial stack, this option makes projections
around the patient. This same operation can alternately be
performed by switching the stacks orientation (for example with
To_Coronal) and then using the "X-Axis" or "Y-Axis" options. The
mean value option with this option projects the pixels using
non-integer weights in order to remove the vertical strips which can
appear at some angles with some images. The other
Axis of Rotation options retain the original nearest neighbor algorithm.
3. The inverted look up table message has been removed. In
some cases projecting a stack with an inverted look up table makes
sense; however, the user should be aware that inverting an image can
produce
unexpected projection values. If this PlugInFilter is installed
with argument "LUT_WARNING" (or any argument with "warn" in it), the
warning
message is restored. The LUT on the projected stack is inverted
to
match the input stack only when there is no message. Thus,
installing
with the "LUT_WARNING" option produces the same behavior as the Project
class.
4. An option to multiply the mean value by a constant has been
added. In the case where there is sparse data (few pixels with
high value) the mean can be very low. This option allow the user
to compensate for this effect without compressing the number of gray
shades.
5. An option to set the nearest point to the upper bound has been
added. This option is useful for cases where the nearest point
may be subject to varying amounts of partial voluming.
|