Re: Release of Improved Tria Image Processing Program

Hi Heini,

I take your point. Your perspective is that of the professional user
and mine that of the amateur who also has a general interest in how
deconvolution technology advances. There's no way I would buy one of
the high cost packages unless the vendors sensibly took the view that
they could increase revenue by selling single user copies at a low
price to private individuals for non-commercial or amateur academic
use. Bear in mind that such copies are unlikely to be abused in
academic or commercial institutions because they are all too aware
that they can be found out and that, at the least, causes
embarrassment.

However, I think it must be difficult even for professionals such as
you to find out which of the available packages you can afford best
meets your needs. Which brings me back to the idea that comparisons
based on the same photos would give you information on the quality of
deconvolution if not the general ease of use of the software. Vendors
naturally portray their products in the best light, I would do the
same myself if in their position. Yet, there must be several varieties
of deconvolution is use in various labs and the users themselves could
arrange a compare and contrast excercise. That would also keep the
vendors on their toes both with respect to quality and price.

Selwyn





On Oct 31, 9:28 am, heini <buerg...@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
wrote:
Well, I slightly disagree...
Image Processing in ImageJ is cool, I love the macro-options for batch-
processing and the measurement-tools, µ-manager....
But good deconvolution of fluorescent images is hard work. One can
easily spend weeks in creating a good PSF for one good stack. One PSF
per color. New PSF if the slice is thicker. New PSF for other camera/
objective/specimen... I have an old IA64 with 2 Processors and 8GB of
RAM running on linux to do the hard work, on normal desktops I always
run out of memory when processing stacks. This is definitvely the poor
man`s solution.
What a difference to paste images into huygens (for example) and have
them deconvoluted without creating PSFs first in less than days! But
Huygens is too expensive for my institute.
Too bad for Tria it won`t run on WinXP64, I'd liked to try it.

Regards, Heini


.

Relevant Pages

  • Re: PSF-based deconvolution
    ... >> Are you still investigating PSF-based deconvolution? ... >is called a Line Spread Function (LSF). ... >creating a 3D PSF out of a 2D LSF, ... While with true deconvolution, you may use non-symmetrical kernels, ...
    (comp.periphs.scanners)
  • Re: How about that Mars
    ... Deconvolution increases the noise level when the spatial frequencies ... of the point spread function (PSF). ... Other deconvolution hints are 1) do not compress the image in any way ...
    (sci.astro.amateur)
  • Re: Blind deconvolution???
    ... Please excuse me that my answer does not really pertain to the "Blind ... Deconvolution" topic but rather concerns deconvolution with know PSF. ... I used the codes provided at: ...
    (sci.image.processing)
  • Re: Release of Improved Tria Image Processing Program
    ... But good deconvolution of fluorescent images is hard work. ... easily spend weeks in creating a good PSF for one good stack. ... Huygens is too expensive for my institute. ...
    (sci.techniques.microscopy)