PSF

[john ferguson (BI)]

Hello,

I’m wanting to do deconvolution using a measured PSF obtained under the same conditions I will use to do my imaging. To build the point spread function I have taken 30 images of subresolution beads but I am unsure of the best way to combine these images. Are they simply aligned on the x,y and z and then averaged or is there other image processing that occurs at this stage? Does anyone know of any software that can do this? Has anything been written for imagej?

Thanks for your help.

John

[Martin Wessendorf-2]

Dear John--

The psf is the entire stack of images--it's a 3-D image that represents
how light originating from a point source, spreads as a function of your
imaging system.

The ImageJ user group would be able to help you with regard to where you
go from here.  You might also check with SVI, the company that makes the
Huygens program: http://www.svi.nl .  Their site has a lot of basic info
on deconvolution, for instance:
http://www.svi.nl/support/wiki/HuygensDeconvolution .  "No commercial
interest."

Good luck!

Martin Wessendorf

john ferguson (BI) wrote:
> Hello,
>
> I’m wanting to do deconvolution using a measured PSF obtained under the same conditions I will use to do my imaging. To build the point spread function I have taken 30 images of subresolution beads but I am unsure of the best way to combine these images. Are they simply aligned on the x,y and z and then averaged or is there other image processing that occurs at this stage? Does anyone know of any software that can do this? Has anything been written for imagej?
>
> Thanks for your help.
>
> John

--
Martin Wessendorf, Ph.D.                   office: [phone]
Assoc Prof, Dept Neuroscience                 lab: [phone]
University of Minnesota             Preferred FAX: [phone]
6-145 Jackson Hall, 321 Church St. SE    Dept Fax: [phone]
Minneapolis, MN  55455                    e-mail: [hidden email]

[Rietdorf, Jens]

Hello John,

Try http://bigwww.epfl.ch/algorithms/deconvolutionlab/

Cheers, jens

-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]]
On Behalf Of john ferguson (BI)
Sent: Tuesday, January 19, 2010 9:31 PM
To: [hidden email]
Subject: PSF

Hello,

I'm wanting to do deconvolution using a measured PSF obtained under the
same conditions I will use to do my imaging. To build the point spread
function I have taken 30 images of subresolution beads but I am unsure
of the best way to combine these images. Are they simply aligned on the
x,y and z and then averaged or is there other image processing that
occurs at this stage? Does anyone know of any software that can do this?
Has anything been written for imagej?

Thanks for your help.

John

[Jose Viña]

Dear Martin,

thanks for the reference! I think John was saying that he has multiple
3D images from different beads, and suggesting that they should be
aligned in space and averaged somehow...

> You might also check with SVI, the company that makes the
> Huygens program: http://www.svi.nl .

Right, this is a good starting point:
http://support.svi.nl/wiki/PsfDistiller. Guidelines for proper bead
image acquisitions are available here:
http://support.svi.nl/wiki/RecordingBeads.

Regards,

jose.

[Ian Dobbie]

"john ferguson (BI)" <[hidden email]> writes:

> Hello,
>
> I’m wanting to do deconvolution using a measured PSF obtained under
> the same conditions I will use to do my imaging. To build the point
> spread function I have taken 30 images of subresolution beads but I am
> unsure of the best way to combine these images. Are they simply
> aligned on the x,y and z and then averaged or is there other image
> processing that occurs at this stage? Does anyone know of any software
> that can do this? Has anything been written for imagej?

Its not exactly clear what your 30 images are. Are they 30 3-D stacks of
single beads? or is it a stack of 30 images of a single bead?

You need a 3D stack to do deconvolution. It is possible to use multiple
stacks to improve signal to noise in the resultant PSF but this requires
taking great care in adding the stacks. They must first be aligned to
much better than the resolution to make sure you don't just blur the PSF
and degrade the resolution of future deconvolution. I'm not sure if
ImageJ plugins are available to add multiple beads stacks. I would be
pretty easy to do this by hand by using a 3D center of mass finding
plugin to generate the bead centres, after background subtraction.
Shifting the images so these all lie at some sensible point like
128,128,128 (assuming you have a 256,256,256 volume) and then simply
adding the images. Both the centre of mass and the shifts would have to
be at sub-pixel resolution.

Ian

[john ferguson (BI)]

Thanks for all your replies,

I seem to have caused a bit of confusion with my explanation, I have 30 stacks of single beads that I want to turn into a point spread function.

We are using velocity to do the deconvolution at the moment but it doesn't have a function to average several bead stacks. I was hoping to find something (free) to make this averaged point spread function.

I have tried the Huygens software and while it is good I'm not sure if we will be able to buy it at the moment.

John


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Ian Dobbie
Sent: 19 January 2010 22:34
To: [hidden email]
Subject: Re: PSF

"john ferguson (BI)" <[hidden email]> writes:

> Hello,
>
> I’m wanting to do deconvolution using a measured PSF obtained under
> the same conditions I will use to do my imaging. To build the point
> spread function I have taken 30 images of subresolution beads but I am
> unsure of the best way to combine these images. Are they simply
> aligned on the x,y and z and then averaged or is there other image
> processing that occurs at this stage? Does anyone know of any software
> that can do this? Has anything been written for imagej?

Its not exactly clear what your 30 images are. Are they 30 3-D stacks of
single beads? or is it a stack of 30 images of a single bead?

You need a 3D stack to do deconvolution. It is possible to use multiple
stacks to improve signal to noise in the resultant PSF but this requires
taking great care in adding the stacks. They must first be aligned to
much better than the resolution to make sure you don't just blur the PSF
and degrade the resolution of future deconvolution. I'm not sure if
ImageJ plugins are available to add multiple beads stacks. I would be
pretty easy to do this by hand by using a 3D center of mass finding
plugin to generate the bead centres, after background subtraction.
Shifting the images so these all lie at some sensible point like
128,128,128 (assuming you have a 256,256,256 volume) and then simply
adding the images. Both the centre of mass and the shifts would have to
be at sub-pixel resolution.

Ian

[Guy Cox]

You might look into the electron microscope area.  It's quiet common in EM to average lots of noisy images on small objects to obtain an overall image.  This requires quite accurate registration, in the presence of substantial noise, which is exactly what you need.  The only thing is that you might have to  do each layer separately since EM doesn't acquire 3D the same way as confocal.  But that shouldn't be too hard.  I can't point you to any specific software, but I bet there is something.

                                Guy

Optical Imaging Techniques in Cell Biology
by Guy Cox    CRC Press / Taylor & Francis
     http://www.guycox.com/optical.htm
______________________________________________
Associate Professor Guy Cox, MA, DPhil(Oxon)
Electron Microscope Unit, Madsen Building F09,
University of Sydney, NSW 2006

Phone [phone]     Fax [phone]
             Mobile [phone]
______________________________________________
      http://www.guycox.net
 


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of john ferguson (BI)
Sent: Thursday, 21 January 2010 5:41 AM
To: [hidden email]
Subject: Re: PSF

Thanks for all your replies,

I seem to have caused a bit of confusion with my explanation, I have 30 stacks of single beads that I want to turn into a point spread function.

We are using velocity to do the deconvolution at the moment but it doesn't have a function to average several bead stacks. I was hoping to find something (free) to make this averaged point spread function.

I have tried the Huygens software and while it is good I'm not sure if we will be able to buy it at the moment.

John


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Ian Dobbie
Sent: 19 January 2010 22:34
To: [hidden email]
Subject: Re: PSF

"john ferguson (BI)" <[hidden email]> writes:

> Hello,
>
> I’m wanting to do deconvolution using a measured PSF obtained under
> the same conditions I will use to do my imaging. To build the point
> spread function I have taken 30 images of subresolution beads but I am
> unsure of the best way to combine these images. Are they simply
> aligned on the x,y and z and then averaged or is there other image
> processing that occurs at this stage? Does anyone know of any software
> that can do this? Has anything been written for imagej?

Its not exactly clear what your 30 images are. Are they 30 3-D stacks of
single beads? or is it a stack of 30 images of a single bead?

You need a 3D stack to do deconvolution. It is possible to use multiple
stacks to improve signal to noise in the resultant PSF but this requires
taking great care in adding the stacks. They must first be aligned to
much better than the resolution to make sure you don't just blur the PSF
and degrade the resolution of future deconvolution. I'm not sure if
ImageJ plugins are available to add multiple beads stacks. I would be
pretty easy to do this by hand by using a 3D center of mass finding
plugin to generate the bead centres, after background subtraction.
Shifting the images so these all lie at some sensible point like
128,128,128 (assuming you have a 256,256,256 volume) and then simply
adding the images. Both the centre of mass and the shifts would have to
be at sub-pixel resolution.

Ian

No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 9.0.730 / Virus Database: 270.14.122/2590 - Release Date: 01/20/10 04:49:00