 |
|
| Computers Forum Index » Computer - Graphics » QuickDraw Regions for efficient ROI description |
|
Page 1 of 1 |
|
| Author |
Message |
| Lawson English |
 Posted: Mon Apr 16, 2007 3:32 am |
|
|
|
Guest
|
A thought has occurred to me...
Many years ago, Apple computer patented the QuickDraw Region algorithm.
http://www.mactech.com/articles/mactech/Vol.05/05.06/ConvertPict2Rgn/index.html
While I understand that many algorithms have been used since then to
efficiently implement the same thing, I'm wondering if anyone has ever
looked at extending the QD Region strategy to 32 bits(or more) for use
in medical/scientific imaging software/hardware?
The original algorithm was designed for bitmapped images with less than
32K total pixels, but the algorithm appears to be extensible.
The reason why I am asking is because I've been dealing with OsiriX
lately, specifically with Regions of Interest (ROIs). As I understand
it, a 3D ROI is a 3D matrix of B&W pixels. It seems to me that an
extended QD Region (EQD) specification could be a VERY easy way to store
a large 3D ROI. You'd just store an array of EQD regions, 1 region per
2D layer of voxels.
A strategy to take full advantage of the potential space savings would
be to prepend the 3D ROI description with a 0, 1, 2 0r 3. A "0"
indicates that the ROI is not a QD Region The other options specify
which orientation of the 3D bounding box is used, since 3D QD regions
could be oriented in any of 3 ways, depending on which is most efficient
(in this case, the way you slice it DOES matter).
2D QD regions also allow for efficient boolean algebra to be applied on
the region data. I assume this could be extended to 3D as well which
would make union, difference, intersection and xoring quite efficient
for 3D ROIs.
In case you were wondering, the QD Region patent expired several years
ago, but apparently no-one has looked at using it for anything since
then. It seems a waste, to me. |
|
|
| Back to top |
|
|
|
|
|
All times are GMT
The time now is Thu Nov 26, 2009 4:24 pm
|
|