| Ongoing Object Graphics Quest [message #28026] |
Sun, 18 November 2001 17:23  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Folks,
I've been on a bit of a quest lately to learn
more about object graphics. I continue to be
amazed at how hard it is to figure everything
out. Partly this is because object graphics are
so haughty. When you make an absolutely asinine
mistake, they just stare at you with this incredibly
disgusted look on their face. They don't say *anything*,
unlike direct graphics, which sometimes deign to
send you a cryptic error message that can occasionally
get you back on track.
We spend an inordinate amount of time just
staring at each other, although I usually have
more of a blank look on my face than an intimidating
one. :-(
Anyway, to solve a problem for a client, and to
give myself a challenging programming exercise
I decided to work on an image processing application
that would allow me to interactively set the
contrast/brightness (also called the window/level)
of an image. Several weeks ago I reported on a
similar program I had written in direct graphics.
(That direct graphics program, WindowImage, has
been upgraded, incidentally, as a result of lessons
learned in the past couple of days.)
Because I has already *mostly* solved the contrast/
brightness problem and I (naively) believed that
converting that to object graphics wouldn't be much
of a challenge, I decided to make the problem more
difficult. I also wanted to be able to zoom the image
"in place", and in a way that preserved the aspect
ratio of the zoomed image subset. This was a
bigger challenge because laying things out in
object graphics windows (at least according to all
the examples RSI provides and my own experience)
is one gigantic pain in the ol' wazoo. I wanted to
develop a rational way of doing this that I could
explain to someone.
Finally, I wanted to know how to have several
"views" of data in one graphics window, and how
to interact with those views independently. (Had
I thought about this for longer than five minutes I would
certainly have given the whole project up as hopeless
before I went to all this trouble.)
The result is a new program on my web page, named
ContrastZoom.
http://www.dfanning.com/programs/contrastzoom.pro
There are three "views" in the window. The view on
the left is the zoom window. You can draw a rubberband
box about a portion of the image that you want to
see closer up. Although I don't zoom into the image,
I take that portion and display it in the same location
in the graphics window in a way that preserves its
aspect ratio. I call this "zoom in place", because
the effect is to zoom into a particular location. You
can go back to the entire image by just clicking and
releasing the cursor in that window.
The center "view" or image is the image that you
use to adjust the contrast/brightness of the
image. Dragging the cursor horizontally sets the
brightness or level. Dragging the cursor vertically
sets the contrast or window. Clicking and releasing
will set the original values of 25% contrast and
75% brightness.
What I particularly like about this program is that
the colorbar in the third "view" on the right
reflects the current window and level. I know
I will offend you medical guys with some color,
but the rest of us can see this better by running
the program with a red-temperature color table like this:
IDL>ContrastZoom, Colortable=3
I understand that this is not the best program I've
even written. But I spent most of 10 days writing it,
and I think even as it is, it might save someone else
a heck of a lot of time. I already have ideas for how
the program can be improved if I decide to put the
lessons learned here in a book.
(By the way, I didn't get anyone responding with
suggestions for improving my previous contrast/brightness
algorithm, so I had to do it myself. I'm still not
totally in love with it, but it's getting better
every time I work on it. At least with the color bar
feedback, I can tell it works now the way I expect
it to work. That's something, anyway.)
As always, I appreciate the feedback.
If you haven't visited my web page in a while,
you can find several new programs at the usual place:
http://www.dfanning.com/documents/programs.html
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
