Re: Suggest method



On Apr 16, 5:23 am, Pera.Maj...@xxxxxxxxx wrote:
Thank you guys for replay very much,

both ideas look very good! But if you look beter on picture you can
see some splashes of color behind of blue or red dots. I need to
filter them somehow before start counting.

But, because this is student project I need some "fancy" stuff, some
mathematic. Not simple (but working algorithm) but something like FFT,
Spatial or Freq analyses, maybe quaternion method...

Bye!

OK pera, pretty funny. I loved aruzinsky's response also. This looks
pretty simple - just find some thresholds in a color space, such as
RGB, threshold it, find connected components, and count them. If you
want to fancy it up some (which might not be necessary), you could
always run the image through a median filter first, then convert the
image into LAB color space and threshold that instead of the RGB
image. Then count blobs and filter them based on size (throw out
blobs too big or too small). Then you could import some astronomical
images from the Digitzed Sky Survey (http://archive.stsci.edu/dss/)
and compare the pattern of the centroids of your blobs to the space
photos to see what part of space (i.e. a constellation) best matches
your spot pattern (use the Groth algorithm for this). Then you could
take a screenshot of the 3D color gamut of the image using this plug-
in (http://rsb.info.nih.gov/ij/plugins/color-inspector.html) for free
image analysis software ImageJ.

Finally for some aruzinskian fun, you could use the Goldbergian twist
on the Las Vegas technique- a sophisticated variant of the Monte Carlo
method ;-) Then compute the "degree of bogosity" (http://catb.org/
jargon/html/B/bogosity.html) and make a scatterplot of this against
the increase in the thickness of the bozone layer (http://
www.urbandictionary.com/define.php?term=Bozone) Now, cite GoogleLabs
to show that the number of web searches for "homework" is directly
correlated with the number of web searches for "imaging" (http://
tinyurl.com/6ftoh5), and infer from that that most homework in the
world involves imaging.

Finally wrap up with a discussion of Quantum Bogodynamics (http://
catb.org/jargon/html/Q/quantum-bogodynamics.html) And of course you'd
want to publish your findings in the Journal of Irreproducible Results
(http://www.jir.com) and reap fame and glory.
Regards,
ImageAnalyst
.



Relevant Pages

  • Re: AVI Decompressor
    ... I am building a capture graph using a XFX GeForce 7 Series ... behavior with the AVI Decompressor filter. ... The card WDM driver only supports the UYVY color space on the capture ... an AVI Decompressor is forced in the graph (with the VMR7 a Color Space ...
    (microsoft.public.win32.programmer.directx.video)
  • Re: AVI Decompressor
    ... I am building a capture graph using a XFX GeForce 7 Series ... behavior with the AVI Decompressor filter. ... The card WDM driver only supports the UYVY color space on the capture ... an AVI Decompressor is forced in the graph (with the VMR7 a Color Space ...
    (microsoft.public.win32.programmer.directx.video)
  • Re: CTransform+InfiniteFilter dont connect to Renderer
    ... especially with the tee filter. ... the renderer and see if that helps. ... The color space converter will use ...
    (microsoft.public.win32.programmer.directx.video)
  • Failing to connect to VMR9 - please advise
    ... I've added my own source filter and the VMR9 to a graph. ... VMR9's input pin's allocator, ... I can connect via a Color Space Converter filter (either by adding ...
    (microsoft.public.win32.programmer.directx.video)
  • Failing to connect to VMR9 - please advise
    ... I've added my own source filter and the VMR9 to a graph. ... VMR9's input pin's allocator, ... I can connect via a Color Space Converter filter (either by adding ...
    (microsoft.public.multimedia.directx.dshow.programming)