Re: Colocalization Test Plugin - 2D images - problem

[Daniel James White]

Hi Oli,

seems like 3D stacks work as expected, but 2D images have a problem...
it takes a very long time for each Costes test iteration...
or maybe its even stuck in a loop.

in any case, this plugin uses a white noise image for the costes test,
which is not what is written in the Costes paper,
which says:  shuffle the info in the image1 and use that as the test image against image 2.

the new colo2 plugin does do the right thing...

rather than try to fix the hard to read and pretty ugly code of the old plugin,
we will work on getting the new coloc2 plugin into the Fiji updater
and then depreciate the old coloc plugins, as obsolete and broken and buggy.

cheers
D

On Mar 12, 2011, at 8:45 AM, Dan White wrote:

> Hi Oli,
>
> On Mar 11, 2011, at 10:21 PM, [hidden email] wrote:
>
>> Colocalization Plugin fails to provide Rcoloc value
>> Olivier <[hidden email]> Mar 11 01:06AM -0800 ^
>>
>> And on another note, while trying the Colocalization Test plugin with
>> the Costes Approximation method, the machine takes over 10 minutes to
>> calculate one iteration, which seems quite odd. The image I tried it
>> on is about 300x200 pixels and on several machines.
>
> that does indeed seem slower than normal.
>
> can you send me the image file so i can figure out why?
>
>> Thanks again for your time
>
> I am testing it now, and something seems to be wrong... it did not used to be this slow.
>
> Fay method still works fast, but Costes is running very slowly.... or even is stuck in a loop.
>
> We have written a new much improved coloc plugin that does not approximate the Costes
> test with a white noise image.... we think that is likely not really the right thing to do....
> but really does scramble the image in blocks of the size of the PSF.
>
> We are verifying that recent changes to imglib have not broken it,
> then we will release it to the Fiji updater....
>
> do you want to help us test that, and improve it so it works the way you want?
>
> see here for more info
> https://docs.google.com/View?id=df66rgc7_2dtqkv3dx
>
> Dan
>
>>
>> Oli
>>
>>
>> Olivier <[hidden email]> Mar 11 01:12AM -0800 ^
>>
>> Last comment, then I'll stop :)
>>
>> Just found out that if I change the bit depth from 16-bit to 8-bit the
>> program runs just fine and quickly enough.
>> The Rcoloc value also appears properly for any image size PROVIDED the
>> image is 8-bit!
>
>
>
>>
>> Is there really such an extra caculation time for 16-bit images or is
>> there something else?
>>
>> Thank you!
>>
>> Oli
>
> Dr. Daniel James White BSc. (Hons.) PhD
> Senior Microscopist / Image Processing and Analysis
> Light Microscopy Facility
> Max Planck Institute of Molecular Cell Biology and Genetics
> Pfotenhauerstrasse 108
> 01307 DRESDEN
> Germany
>
> [phone] (German Mobile)
> +49  (0)351 210 2627 (Work phone at MPI-CBG)
> +49  (0)351 210 1078 (Fax MPI-CBG LMF)
>
> http://www.bioimagexd.net                BioImageXD
> http://pacific.mpi-cbg.de                Fiji (is just ImageJ - batteries included)
> http://www.chalkie.org.uk
> [hidden email]
> ( [hidden email] )
>
>
>
>
>
>

Dr. Daniel James White BSc. (Hons.) PhD
Senior Microscopist / Image Processing and Analysis
Light Microscopy Facility
Max Planck Institute of Molecular Cell Biology and Genetics
Pfotenhauerstrasse 108
01307 DRESDEN
Germany

[phone] (German Mobile)
+49  (0)351 210 2627 (Work phone at MPI-CBG)
+49  (0)351 210 1078 (Fax MPI-CBG LMF)

http://www.bioimagexd.net                BioImageXD
http://pacific.mpi-cbg.de                Fiji (is just ImageJ - batteries included)
http://www.chalkie.org.uk
[hidden email]
( [hidden email] )