32bit colocalization plugin

[Colin Rickman]

Hi,

I currently use the colocalization threshold plugin to calculate the
linear regression between two channels. At present this plugin can only
use 8bit or 16bit unsigned byte images. Is it possible to change the
plugin to accept 32bit float images or is there another plugin out there
which can do this.

Hoe you can help

Colin

--
Dr Colin Rickman
Centre for Integrative Physiology
School of Biomedical Sciences
University of Edinburgh
Hugh Robson Building
George Square
Edinburgh
EH8 9XD

Tel: [phone]
Fax: [phone]


The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

[Tony Collins-4]

Hi Colin,

May be tricky. The original algorithm steps through the data setting a threshold at "intensity/2" until it finds one where the Pearson's value is zero. With a 32-bit image this would require too many steps!

I'm also not sure 32-bit would give you much more information. Pearson's coefficient normalises the data anyway so you may want to scale it to 16-bits.

Hope this helps,

Tony

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Colin Rickman
Sent: November-03-08 6:17 AM
To: [hidden email]
Subject: 32bit colocalization plugin

Hi,

I currently use the colocalization threshold plugin to calculate the
linear regression between two channels. At present this plugin can only
use 8bit or 16bit unsigned byte images. Is it possible to change the
plugin to accept 32bit float images or is there another plugin out there
which can do this.

Hoe you can help

Colin

--
Dr Colin Rickman
Centre for Integrative Physiology
School of Biomedical Sciences
University of Edinburgh
Hugh Robson Building
George Square
Edinburgh
EH8 9XD

Tel: [phone]
Fax: [phone]


The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

[Colin Rickman]

Hi Tony,

Thanks for getting back to me. I'm not after the threshold values. I use
the values for the gradient (m) and y-intercept (b) given by the plugin
to calculate the residual for each pixel in the image (which I then
colour code).  The calculations for this all use 32 bit float (to permit
both negative and non-integer values). I am trying to avoid converting
from 32bit (the data currently contains non-integer values) to signed
integer to use the plugin but then calculating the residuals back in
32bit float.

If there is an alternative plugin which can provide the gradient,
y-intercept and Pearsons correlation coefficient using two 32bit float
3D stacks that would be great.

Colin

Tony Collins wrote:

> Hi Colin,
>
> May be tricky. The original algorithm steps through the data setting a threshold at "intensity/2" until it finds one where the Pearson's value is zero. With a 32-bit image this would require too many steps!
>
> I'm also not sure 32-bit would give you much more information. Pearson's coefficient normalises the data anyway so you may want to scale it to 16-bits.
>
> Hope this helps,
>
> Tony
>
> -----Original Message-----
> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Colin Rickman
> Sent: November-03-08 6:17 AM
> To: [hidden email]
> Subject: 32bit colocalization plugin
>
> Hi,
>
> I currently use the colocalization threshold plugin to calculate the
> linear regression between two channels. At present this plugin can only
> use 8bit or 16bit unsigned byte images. Is it possible to change the
> plugin to accept 32bit float images or is there another plugin out there
> which can do this.
>
> Hoe you can help
>
> Colin
>
> --
> Dr Colin Rickman
> Centre for Integrative Physiology
> School of Biomedical Sciences
> University of Edinburgh
> Hugh Robson Building
> George Square
> Edinburgh
> EH8 9XD
>
> Tel: [phone]
> Fax: [phone]
>
>
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>  

--
Dr Colin Rickman
Centre for Integrative Physiology
School of Biomedical Sciences
University of Edinburgh
Hugh Robson Building
George Square
Edinburgh
EH8 9XD

Tel: [phone]
Fax: [phone]


The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.