My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives...
Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? Thanks for educating me if you can! John Oreopoulos Research Assistant Spectral Applied Research Richmond Hill, Ontario Canada |
A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very bright edge artifacts in the ratio images. So I checked the system with 0.2 um beads in the same mounting media. I found that shifting the focus 0.3 um for CFP from the YFP image solved the problem.
We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. The vendor assisted us and we found that by sandwiching a 1000 mm lens in with one of the filters, the focus was in registration. I don't see how we could have solved either of these focus issues post image collection. ________________________________________________________ Michael Cammer, Assistant Research Scientist Skirball Institute of Biomolecular Medicine Lab: (212) 263-3208 Cell: (914) 309-3270 -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John Oreopoulos Sent: Monday, October 17, 2011 9:26 PM To: [hidden email] Subject: Plugin to correct for objective axial chromatic aberration? My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives... Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? Thanks for educating me if you can! John Oreopoulos Research Assistant Spectral Applied Research Richmond Hill, Ontario Canada ------------------------------------------------------------ This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain information that is proprietary, confidential, and exempt from disclosure under applicable law. Any unauthorized review, use, disclosure, or distribution is prohibited. If you have received this email in error please notify the sender by return email and delete the original message. Please note, the recipient should check this email and any attachments for the presence of viruses. The organization accepts no liability for any damage caused by any virus transmitted by this email. ================================= |
John,
This is indeed, an important issue, since many people do not consider this when doing colocalization studies. I assume that your data come from a z-series in which you are recording each channel separately. Moreover, I assume that your z-series has enough resolution to allow you to determine the number of slices that would be necessary to remove in order to line up the center of each image. It should then be possible to manually remove the appropriate number of slices from the end of the stack so that, say, if the red z-stack goes from 1-20, the blue stack goes from 3-20, etc. You would then have to truncate the red stack so that the number of slices were even, so that the red stack should be 1-18 or so. You could then merge the stacks back together. --or were you looking for something more elaborate? Might there be a magnification issue as well? In my experience, we don't see that. Joel On Wed, Oct 19, 2011 at 9:44 AM, Cammer, Michael <[hidden email] > wrote: > A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X > N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very > bright edge artifacts in the ratio images. So I checked the system with 0.2 > um beads in the same mounting media. I found that shifting the focus 0.3 um > for CFP from the YFP image solved the problem. > > We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. > The vendor assisted us and we found that by sandwiching a 1000 mm lens in > with one of the filters, the focus was in registration. > > I don't see how we could have solved either of these focus issues post > image collection. > > ________________________________________________________ > Michael Cammer, Assistant Research Scientist > Skirball Institute of Biomolecular Medicine > Lab: (212) 263-3208 Cell: (914) 309-3270 > > -----Original Message----- > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John > Oreopoulos > Sent: Monday, October 17, 2011 9:26 PM > To: [hidden email] > Subject: Plugin to correct for objective axial chromatic aberration? > > My apologies as I suspect this question may have been posed to the > listserver before, but I could not find an answer in the discussion > archives... > > Suppose I have a well aligned microscope system and have made every effort > to minimize refractive index mistmatch, etc. Further suppose then that I > perform PSF measurements on some multi-colored sub-diffraction sized > fluorescent microbeads in two or three different spectral channels. In this > very unfortunate case, finally suppose that I measure an axial chromatic > shift of the beads in the different spectral channels (even it may be an > apochromat objective). > > My question is, once I know the amount of axial shift in each channel for > this particular objective lens, is there a way to correct and realign images > in the z-direction through image processing post-acqisution? Has anyone > developed an ImageJ or FIJI plugin to do that? > > I would think that this is a very important aspect of colocalization (since > colocalization is really a 3D problem, not just 2D), or other ratiometric > fluorescence measurements such as FRET or calcium imaging. Surely the > solution is not just to get a better chormatically corrected objective, is > it? My sense is that this would be a problem that the experts in > deconvolution fluorescence microscopy have already tackled as well. Is that > the case? Is there any literature on this topic? > > Thanks for educating me if you can! > > John Oreopoulos > Research Assistant > Spectral Applied Research > Richmond Hill, Ontario > Canada > > ------------------------------------------------------------ > This email message, including any attachments, is for the sole use of the > intended recipient(s) and may contain information that is proprietary, > confidential, and exempt from disclosure under applicable law. Any > unauthorized review, use, disclosure, or distribution is prohibited. If you > have received this email in error please notify the sender by return email > and delete the original message. Please note, the recipient should check > this email and any attachments for the presence of viruses. The organization > accepts no liability for any damage caused by any virus transmitted by this > email. > ================================= > -- Joel B. Sheffield, Ph.D Department of Biology Temple University Philadelphia, PA 19122 Voice: 215 204 8839 e-mail: [hidden email] URL: http://astro.temple.edu/~jbs |
Hi,
On Wed, Oct 19, 2011 at 17:30, JOEL B. SHEFFIELD <[hidden email]> wrote: > You could then merge the stacks back together. --or were > you looking for something more elaborate? Might there be a magnification > issue as well? In my experience, we don't see that. > > Joel > > Your suggestion of shifting the stacks seems to be the most sensible solution. As regards magnification, I thought objectives were compensated for that, until I experimented strange shifts using a Leica confocal (the "benchtop" TC-SPE). We had one with a 405 nm laser and both UV-optimized and non-UV objectives. Registration of the channels was OK with the UV objectives, but not with the non-UV ones, with a difference in magnification for the different wavelengths. According to the people at Leica, difference in magnification between wavelength is corrected in the microscope tube, and is different for UV or non-UV lasers & objectives, and our scope was optimized for the UV configuration. To be able to use non-UV objectives (that had better NA), I ended up mapping the difference in magnification (and also a slight translational shift) using a dense Tetraspeck beads field and wrote a macro that expanded or shrunk the channels to register them in agreement with the measured differences. I can send the macro if someone is interested. Christophe > On Wed, Oct 19, 2011 at 9:44 AM, Cammer, Michael < > [hidden email] > > wrote: > > > A few years ago we were using a CFP-YFP biosensor in fixed cells with a > 60X > > N.A. 1.4 Olympus objective on an IX 70. We found that we were getting > very > > bright edge artifacts in the ratio images. So I checked the system with > 0.2 > > um beads in the same mounting media. I found that shifting the focus 0.3 > um > > for CFP from the YFP image solved the problem. > > > > We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. > > The vendor assisted us and we found that by sandwiching a 1000 mm lens > in > > with one of the filters, the focus was in registration. > > > > I don't see how we could have solved either of these focus issues post > > image collection. > > > > ________________________________________________________ > > Michael Cammer, Assistant Research Scientist > > Skirball Institute of Biomolecular Medicine > > Lab: (212) 263-3208 Cell: (914) 309-3270 > > > > -----Original Message----- > > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of > John > > Oreopoulos > > Sent: Monday, October 17, 2011 9:26 PM > > To: [hidden email] > > Subject: Plugin to correct for objective axial chromatic aberration? > > > > My apologies as I suspect this question may have been posed to the > > listserver before, but I could not find an answer in the discussion > > archives... > > > > Suppose I have a well aligned microscope system and have made every > effort > > to minimize refractive index mistmatch, etc. Further suppose then that I > > perform PSF measurements on some multi-colored sub-diffraction sized > > fluorescent microbeads in two or three different spectral channels. In > this > > very unfortunate case, finally suppose that I measure an axial chromatic > > shift of the beads in the different spectral channels (even it may be an > > apochromat objective). > > > > My question is, once I know the amount of axial shift in each channel for > > this particular objective lens, is there a way to correct and realign > images > > in the z-direction through image processing post-acqisution? Has anyone > > developed an ImageJ or FIJI plugin to do that? > > > > I would think that this is a very important aspect of colocalization > (since > > colocalization is really a 3D problem, not just 2D), or other ratiometric > > fluorescence measurements such as FRET or calcium imaging. Surely the > > solution is not just to get a better chormatically corrected objective, > is > > it? My sense is that this would be a problem that the experts in > > deconvolution fluorescence microscopy have already tackled as well. Is > that > > the case? Is there any literature on this topic? > > > > Thanks for educating me if you can! > > > > John Oreopoulos > > Research Assistant > > Spectral Applied Research > > Richmond Hill, Ontario > > Canada > > > > ------------------------------------------------------------ > > This email message, including any attachments, is for the sole use of the > > intended recipient(s) and may contain information that is proprietary, > > confidential, and exempt from disclosure under applicable law. Any > > unauthorized review, use, disclosure, or distribution is prohibited. If > you > > have received this email in error please notify the sender by return > > and delete the original message. Please note, the recipient should check > > this email and any attachments for the presence of viruses. The > organization > > accepts no liability for any damage caused by any virus transmitted by > this > > email. > > ================================= > > > > > > -- > > > Joel B. Sheffield, Ph.D > Department of Biology > Temple University > Philadelphia, PA 19122 > Voice: 215 204 8839 > e-mail: [hidden email] > URL: http://astro.temple.edu/~jbs > |
In reply to this post by John Oreopoulos
Hi John and Michael,
From a colocalization perspective.... Objective lenses are never perfect, not even really expensive ones that claim multi color chromatic correction. When it comes to imaging diffraction limited objects with a 1.4 NA lens on a widefield system there will still be a unique colour shift in all three spatial dimensions for that particular lens, and even how it is screwed in. Also the alignment of the fluorescence filters in different cubes might bend different channels to slightly different places on the detector. On the OMX you can even adjust this... normally its fixed on most systems but might also be tweakable. Olympus does all the chromatic correction in the lens, but others do some in the tube lens. This is why its not a good idea to screw a Zeiss c-apochromat 100x 1.6 oil into an olympus stand. So we are left with a systematic error that can be measured, actually to a precision of 10s of nm by Gaussian fitting bead images (like you do in PALM/SSTOR type super resolution and single molecule tracking. ) On a single point scanning confocal the matter is made worse by the more complicated optics. The Zeiss 510 has a different pinhole per channel... and depending on their positions and correct setup, the images of the bead in different channels is also affected by pinhole position settings. Nowadays with the Zeiss 7xx series and on the Olympus FV1000 and other modern point scanners there is only 1 pinhole for the emission light, so that problem is suppressed. BUT, the 405 lasers come in through a different fibre than the Vis lasers.. and so there is a collimation adjustment to make Same for 2photon lasers.... getting the near UV and IR laser spots to coincide exactly with the vis lines is hard... so its best to get is as close as is reasonable to do, then measure the residual error then correct for it. There always remain measurable errors (unless you were really really lucky and happen to have a perfect lens, which is a 1 in a 1000 chance at best, plus perfectly aligned fluorescence filters) OK, so at the simple level, we can assume that the whole field of view is shifted by some 3D shift per channel compared to the reference, usually green channel. So if we measure a few 1 or 0.5 or so micron beads (no need to use tiny beads here) near the centre of the field of view (where optics are best) we can calculate or guestimate the center of mass of the bead images in each channel, and work out the shift vectors needed. These will be between 10-1000 nm in xy and up to 500 or even 2000 micron in z. BUT, they will never be whole xy pixels or z slice shifts... forget that idea its too crude. see the 2 slides titled "Check with multi-colour beads" at http://ifn.mpi-cbg.de/wiki/ifn/images/0/0d/QuantitativeColocAnalysis-12-2010.pdf Now we can use Erik M's TransformJ Translate plugin in Fiji/ImageJ to do a sub pixel resolution shift for for each channel, as we have measured those shifts. http://www.imagescience.org/meijering/software/transformj/translate.html Use a nice interpolation method to avoid smashing the information in your images - eg quintic B-spline interpolation Or one could imaging a Fourier based method... (using a phase shift?) THe 3D stitching plugin contains stuff that might do this but its not exposed in the GUIs... so scripting would be required there. Moving the images using a whole pixel or z plane shift will not be precise enough for high res colocalization analysis. At a higher level there may also be a magnification difference between the different colour channels. On multi camera systems, like OMX or some HCS system there WILL be a rotation angle difference between the channel images. TransformJ Affine plugin might do the job, which can also expand or shrink to e image to accommodate different magnification and rotation. Even worse could be a non linear warp of the image that is different per channel. Again these are measurable and fixable with some effort, eg using bUnwarpJ So depending how precise you need to be, over how large a field of view... the difficulty of the correction varies. We could add this kind of colour shift correction to the new Coloc_2 plugin... by reusing TransformJ and maybe also bUnwarpJ. Its a whole world of fun. cheers Dan On Oct 20, 2011, at 6:02 AM, IMAGEJ automatic digest system wrote: > Date: Wed, 19 Oct 2011 09:44:25 -0400 > From: "Cammer, Michael" <[hidden email]> > Subject: Re: Plugin to correct for objective axial chromatic aberration? > > A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very bright edge artifacts in the ratio images. So I checked the system with 0.2 um beads in the same mounting media. I found that shifting the focus 0.3 um for CFP from the YFP image solved the problem. > > We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. The vendor assisted us and we found that by sandwiching a 1000 mm lens in with one of the filters, the focus was in registration. > > I don't see how we could have solved either of these focus issues post image collection. > > ________________________________________________________ > Michael Cammer, Assistant Research Scientist > Skirball Institute of Biomolecular Medicine > Lab: (212) 263-3208 Cell: (914) 309-3270 > > -----Original Message----- > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John Oreopoulos > Sent: Monday, October 17, 2011 9:26 PM > To: [hidden email] > Subject: Plugin to correct for objective axial chromatic aberration? > > My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives... > > Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). > > My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? > > I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? > > Thanks for educating me if you can! > > John Oreopoulos > Research Assistant > Spectral Applied Research > Richmond Hill, Ontario > Canada Dr. Daniel James White BSc. (Hons.) PhD Leader - Image Processing Facility, Senior Microscopist, Light Microscopy Facility. Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) chalkie666 Skype http://www.bioimagexd.net BioImageXD http://fiji.sc Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
Hi Dan and everyone,
On Thu, Oct 20, 2011 at 09:35, Daniel James White <[hidden email]> wrote: > Hi John and Michael, > > OK, so at the simple level, we can assume that the whole field of view is > shifted by some 3D shift per channel compared to the reference, usually > green channel. So if we measure a few 1 or 0.5 or so micron beads (no need > to use tiny beads here) near the centre of the field of view (where optics > are best) we can calculate or guestimate the center of mass of the bead > images in each channel, and work out the shift vectors needed. These will be > between 10-1000 nm in xy and up to 500 or even 2000 micron in z. > With our Axio Observer widefield scope, the main problem is that from time to time it seems the cubes does not fall exactly into the same place, so something like 1 out of 10 images has a shift clearly different from the others... Even if I force it to wait before taking the image (to make sure the rotating barrel has finished moving/vibrating before taking the image). So ideally you would always embed Tetraspeck beads into your sample to correct shift/warp/whatever transformation you have between your channels independently for each acquired image. Christophe > cheers > > Dan > > > On Oct 20, 2011, at 6:02 AM, IMAGEJ automatic digest system wrote: > > > Date: Wed, 19 Oct 2011 09:44:25 -0400 > > From: "Cammer, Michael" <[hidden email]> > > Subject: Re: Plugin to correct for objective axial chromatic aberration? > > > > A few years ago we were using a CFP-YFP biosensor in fixed cells with a > 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting > very bright edge artifacts in the ratio images. So I checked the system > with 0.2 um beads in the same mounting media. I found that shifting the > focus 0.3 um for CFP from the YFP image solved the problem. > > > > We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. > The vendor assisted us and we found that by sandwiching a 1000 mm lens in > with one of the filters, the focus was in registration. > > > > I don't see how we could have solved either of these focus issues post > image collection. > > > > ________________________________________________________ > > Michael Cammer, Assistant Research Scientist > > Skirball Institute of Biomolecular Medicine > > Lab: (212) 263-3208 Cell: (914) 309-3270 > > > > -----Original Message----- > > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of > John Oreopoulos > > Sent: Monday, October 17, 2011 9:26 PM > > To: [hidden email] > > Subject: Plugin to correct for objective axial chromatic aberration? > > > > My apologies as I suspect this question may have been posed to the > listserver before, but I could not find an answer in the discussion > archives... > > > > Suppose I have a well aligned microscope system and have made every > effort to minimize refractive index mistmatch, etc. Further suppose then > that I perform PSF measurements on some multi-colored sub-diffraction sized > fluorescent microbeads in two or three different spectral channels. In this > very unfortunate case, finally suppose that I measure an axial chromatic > shift of the beads in the different spectral channels (even it may be an > apochromat objective). > > > > My question is, once I know the amount of axial shift in each channel for > this particular objective lens, is there a way to correct and realign images > in the z-direction through image processing post-acqisution? Has anyone > developed an ImageJ or FIJI plugin to do that? > > > > I would think that this is a very important aspect of colocalization > (since colocalization is really a 3D problem, not just 2D), or other > ratiometric fluorescence measurements such as FRET or calcium imaging. > Surely the solution is not just to get a better chormatically corrected > objective, is it? My sense is that this would be a problem that the experts > in deconvolution fluorescence microscopy have already tackled as well. Is > that the case? Is there any literature on this topic? > > > > Thanks for educating me if you can! > > > > John Oreopoulos > > Research Assistant > > Spectral Applied Research > > Richmond Hill, Ontario > > Canada > > Dr. Daniel James White BSc. (Hons.) PhD > > Leader - Image Processing Facility, > Senior Microscopist, > Light Microscopy Facility. > > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > chalkie666 Skype > http://www.bioimagexd.net BioImageXD > http://fiji.sc Fiji - is just ImageJ > (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform > (BioDIP) > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) > |
Can somebody please show some examples of this?
I might have a suggestion that I use for lateral chromatic aberration in brightfield images that I wonder if it might work here too. Cheers Gabriel |
In reply to this post by Daniel James White
Thanks all for the ideas and discussion on this. I didn't realize it was still that big of a problem. Certainly raises the question again of how to properly do a "simple" colocalization experiment and analyze/process the data!
John Oreopoulos On 2011-10-20, at 3:35 AM, Daniel James White wrote: > Hi John and Michael, > > From a colocalization perspective.... > > Objective lenses are never perfect, not even really expensive ones that claim multi color chromatic correction. > When it comes to imaging diffraction limited objects with a 1.4 NA lens on a widefield system > there will still be a unique colour shift in all three spatial dimensions for that particular lens, > and even how it is screwed in. > > Also the alignment of the fluorescence filters in different cubes might bend different channels to slightly different places on the detector. > On the OMX you can even adjust this... normally its fixed on most systems but might also be tweakable. > > Olympus does all the chromatic correction in the lens, but others do some in the tube lens. > This is why its not a good idea to screw a Zeiss c-apochromat 100x 1.6 oil into an olympus stand. > > So we are left with a systematic error that can be measured, actually to a precision of 10s of nm by Gaussian fitting bead images (like you do in PALM/SSTOR type super resolution and single molecule tracking. ) > > On a single point scanning confocal the matter is made worse by the more complicated optics. > The Zeiss 510 has a different pinhole per channel... and depending on their positions and correct setup, > the images of the bead in different channels is also affected by pinhole position settings. > > Nowadays with the Zeiss 7xx series and on the Olympus FV1000 and other modern point scanners there is only 1 pinhole > for the emission light, so that problem is suppressed. > BUT, the 405 lasers come in through a different fibre than the Vis lasers.. and so there is a collimation adjustment to make > Same for 2photon lasers.... getting the near UV and IR laser spots to coincide exactly with the vis lines is hard... so its best to get is as close as is reasonable to do, then measure the residual error then correct for it. > > There always remain measurable errors (unless you were really really lucky and happen to have a perfect lens, which is a 1 in a 1000 chance at best, plus perfectly aligned fluorescence filters) > > OK, so at the simple level, we can assume that the whole field of view is shifted by some 3D shift per channel compared to the reference, usually green channel. So if we measure a few 1 or 0.5 or so micron beads (no need to use tiny beads here) near the centre of the field of view (where optics are best) we can calculate or guestimate the center of mass of the bead images in each channel, and work out the shift vectors needed. These will be between 10-1000 nm in xy and up to 500 or even 2000 micron in z. > BUT, they will never be whole xy pixels or z slice shifts... forget that idea its too crude. > > see the 2 slides titled > "Check with multi-colour beads" > at http://ifn.mpi-cbg.de/wiki/ifn/images/0/0d/QuantitativeColocAnalysis-12-2010.pdf > > Now we can use Erik M's TransformJ Translate plugin in Fiji/ImageJ to do a sub pixel resolution shift for for each channel, as we have measured those shifts. > http://www.imagescience.org/meijering/software/transformj/translate.html > Use a nice interpolation method to avoid smashing the information in your images - eg quintic B-spline interpolation > Or one could imaging a Fourier based method... (using a phase shift?) THe 3D stitching plugin contains stuff that might do this but its not exposed in the GUIs... so scripting would be required there. > > Moving the images using a whole pixel or z plane shift will not be precise enough for high res colocalization analysis. > > At a higher level there may also be a magnification difference between the different colour channels. > On multi camera systems, like OMX or some HCS system there WILL be a rotation angle difference between the channel images. > TransformJ Affine plugin might do the job, which can also expand or shrink to e image to accommodate different magnification and rotation. > > Even worse could be a non linear warp of the image that is different per channel. > Again these are measurable and fixable with some effort, eg using bUnwarpJ > > So depending how precise you need to be, over how large a field of view... the difficulty of the correction varies. > > We could add this kind of colour shift correction to the new Coloc_2 plugin... by reusing TransformJ and maybe also bUnwarpJ. > > Its a whole world of fun. > > cheers > > Dan > > > > On Oct 20, 2011, at 6:02 AM, IMAGEJ automatic digest system wrote: > >> Date: Wed, 19 Oct 2011 09:44:25 -0400 >> From: "Cammer, Michael" <[hidden email]> >> Subject: Re: Plugin to correct for objective axial chromatic aberration? >> >> A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very bright edge artifacts in the ratio images. So I checked the system with 0.2 um beads in the same mounting media. I found that shifting the focus 0.3 um for CFP from the YFP image solved the problem. >> >> We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. The vendor assisted us and we found that by sandwiching a 1000 mm lens in with one of the filters, the focus was in registration. >> >> I don't see how we could have solved either of these focus issues post image collection. >> >> ________________________________________________________ >> Michael Cammer, Assistant Research Scientist >> Skirball Institute of Biomolecular Medicine >> Lab: (212) 263-3208 Cell: (914) 309-3270 >> >> -----Original Message----- >> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John Oreopoulos >> Sent: Monday, October 17, 2011 9:26 PM >> To: [hidden email] >> Subject: Plugin to correct for objective axial chromatic aberration? >> >> My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives... >> >> Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). >> >> My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? >> >> I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? >> >> Thanks for educating me if you can! >> >> John Oreopoulos >> Research Assistant >> Spectral Applied Research >> Richmond Hill, Ontario >> Canada > > Dr. Daniel James White BSc. (Hons.) PhD > > Leader - Image Processing Facility, > Senior Microscopist, > Light Microscopy Facility. > > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > chalkie666 Skype > http://www.bioimagexd.net BioImageXD > http://fiji.sc Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
In reply to this post by Daniel James White
Hi Folks,
The author of the amazing SPIM registration / fusion plugins for multi angle z stacks, Stephan Preibisch, just informed me that the SPIM plugins will do the job of calculating the 3D rigid or even affine transform between two or more z stacks of beads - from eg different colour channels and this transform can be appiled to other images easily.... so measure with beads, then apply transform to real images. Its all part of the normal bead based SPIM registration / image fusion workflow - Well done Steffi! A general solution to a problem in SPIM turns out to be super useful in another related area. So this should work with z stack from any 3D microscope doing multiple channels with either multiple cameras (different xy and rotation), view splitters, or just off alignment dichromatic mirror and filters, and also fix the lateral, axial, and magnification parts of the chromatic aberration. We were looking for a solution for the OMX which has 4 cameras that cant be perfectly aligned to sub pixel accuracy, and to take care of the residual chromatic aberration from the objective lens.... and it was under my nose the whole time.... I should have asked Steffi earlier. I have not tried it yet, and Steffi just updated the plugins to new version and I have to play with them and he has to write the new documentation... but it should work well once we figure out any little problems there may or may not be. Fiji to the rescue once more! Dan On Oct 20, 2011, at 9:35 AM, Daniel James White wrote: > Hi John and Michael, > > From a colocalization perspective.... > > Objective lenses are never perfect, not even really expensive ones that claim multi color chromatic correction. > When it comes to imaging diffraction limited objects with a 1.4 NA lens on a widefield system > there will still be a unique colour shift in all three spatial dimensions for that particular lens, > and even how it is screwed in. > > Also the alignment of the fluorescence filters in different cubes might bend different channels to slightly different places on the detector. > On the OMX you can even adjust this... normally its fixed on most systems but might also be tweakable. > > Olympus does all the chromatic correction in the lens, but others do some in the tube lens. > This is why its not a good idea to screw a Zeiss c-apochromat 100x 1.6 oil into an olympus stand. > > So we are left with a systematic error that can be measured, actually to a precision of 10s of nm by Gaussian fitting bead images (like you do in PALM/SSTOR type super resolution and single molecule tracking. ) > > On a single point scanning confocal the matter is made worse by the more complicated optics. > The Zeiss 510 has a different pinhole per channel... and depending on their positions and correct setup, > the images of the bead in different channels is also affected by pinhole position settings. > > Nowadays with the Zeiss 7xx series and on the Olympus FV1000 and other modern point scanners there is only 1 pinhole > for the emission light, so that problem is suppressed. > BUT, the 405 lasers come in through a different fibre than the Vis lasers.. and so there is a collimation adjustment to make > Same for 2photon lasers.... getting the near UV and IR laser spots to coincide exactly with the vis lines is hard... so its best to get is as close as is reasonable to do, then measure the residual error then correct for it. > > There always remain measurable errors (unless you were really really lucky and happen to have a perfect lens, which is a 1 in a 1000 chance at best, plus perfectly aligned fluorescence filters) > > OK, so at the simple level, we can assume that the whole field of view is shifted by some 3D shift per channel compared to the reference, usually green channel. So if we measure a few 1 or 0.5 or so micron beads (no need to use tiny beads here) near the centre of the field of view (where optics are best) we can calculate or guestimate the center of mass of the bead images in each channel, and work out the shift vectors needed. These will be between 10-1000 nm in xy and up to 500 or even 2000 micron in z. > BUT, they will never be whole xy pixels or z slice shifts... forget that idea its too crude. > > see the 2 slides titled > "Check with multi-colour beads" > at http://ifn.mpi-cbg.de/wiki/ifn/images/0/0d/QuantitativeColocAnalysis-12-2010.pdf > > Now we can use Erik M's TransformJ Translate plugin in Fiji/ImageJ to do a sub pixel resolution shift for for each channel, as we have measured those shifts. > http://www.imagescience.org/meijering/software/transformj/translate.html > Use a nice interpolation method to avoid smashing the information in your images - eg quintic B-spline interpolation > Or one could imaging a Fourier based method... (using a phase shift?) THe 3D stitching plugin contains stuff that might do this but its not exposed in the GUIs... so scripting would be required there. > > Moving the images using a whole pixel or z plane shift will not be precise enough for high res colocalization analysis. > > At a higher level there may also be a magnification difference between the different colour channels. > On multi camera systems, like OMX or some HCS system there WILL be a rotation angle difference between the channel images. > TransformJ Affine plugin might do the job, which can also expand or shrink to e image to accommodate different magnification and rotation. > > Even worse could be a non linear warp of the image that is different per channel. > Again these are measurable and fixable with some effort, eg using bUnwarpJ > > So depending how precise you need to be, over how large a field of view... the difficulty of the correction varies. > > We could add this kind of colour shift correction to the new Coloc_2 plugin... by reusing TransformJ and maybe also bUnwarpJ. > > Its a whole world of fun. > > cheers > > Dan > > > > On Oct 20, 2011, at 6:02 AM, IMAGEJ automatic digest system wrote: > >> Date: Wed, 19 Oct 2011 09:44:25 -0400 >> From: "Cammer, Michael" <[hidden email]> >> Subject: Re: Plugin to correct for objective axial chromatic aberration? >> >> A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very bright edge artifacts in the ratio images. So I checked the system with 0.2 um beads in the same mounting media. I found that shifting the focus 0.3 um for CFP from the YFP image solved the problem. >> >> We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. The vendor assisted us and we found that by sandwiching a 1000 mm lens in with one of the filters, the focus was in registration. >> >> I don't see how we could have solved either of these focus issues post image collection. >> >> ________________________________________________________ >> Michael Cammer, Assistant Research Scientist >> Skirball Institute of Biomolecular Medicine >> Lab: (212) 263-3208 Cell: (914) 309-3270 >> >> -----Original Message----- >> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John Oreopoulos >> Sent: Monday, October 17, 2011 9:26 PM >> To: [hidden email] >> Subject: Plugin to correct for objective axial chromatic aberration? >> >> My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives... >> >> Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). >> >> My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? >> >> I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? >> >> Thanks for educating me if you can! >> >> John Oreopoulos >> Research Assistant >> Spectral Applied Research >> Richmond Hill, Ontario >> Canada > > Dr. Daniel James White BSc. (Hons.) PhD > > Leader - Image Processing Facility, > Senior Microscopist, > Light Microscopy Facility. > > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > chalkie666 Skype > http://www.bioimagexd.net BioImageXD > http://fiji.sc Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) > > > > > > > > > > > > > > > > > Dr. Daniel James White BSc. (Hons.) PhD Leader - Image Processing Facility, Senior Microscopist, Light Microscopy Facility. Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) chalkie666 Skype http://www.bioimagexd.net BioImageXD http://fiji.sc Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
Just to be clear Dan, you mean this plugin, correct?
http://pacific.mpi-cbg.de/wiki/index.php/SPIM_Registration This is indeed very useful information then. Thanks for updating all of us. Hopefully I'll get a chance to try it out. John Oreopoulos On 2011-10-21, at 10:37 AM, Daniel James White wrote: > Hi Folks, > > The author of the amazing SPIM registration / fusion plugins for multi angle z stacks, Stephan Preibisch, > just informed me that the SPIM plugins will do the job of calculating the > 3D rigid or even affine transform between two or more z stacks of beads - from eg different colour channels > and this transform can be appiled to other images easily.... so measure with beads, then apply transform to real images. > > Its all part of the normal bead based SPIM registration / image fusion workflow - Well done Steffi! > A general solution to a problem in SPIM turns out to be super useful in another related area. > > So this should work with z stack from any 3D microscope doing multiple channels > with either multiple cameras (different xy and rotation), view splitters, or just off alignment dichromatic mirror and filters, > and also fix the lateral, axial, and magnification parts of the chromatic aberration. > > We were looking for a solution for the OMX which has 4 cameras that cant be perfectly aligned to sub pixel accuracy, > and to take care of the residual chromatic aberration from the objective lens.... > and it was under my nose the whole time.... I should have asked Steffi earlier. > > I have not tried it yet, and Steffi just updated the plugins to new version and I have to play with them > and he has to write the new documentation... > but it should work well once we figure out any little problems there may or may not be. > > Fiji to the rescue once more! > > Dan > > > > On Oct 20, 2011, at 9:35 AM, Daniel James White wrote: > >> Hi John and Michael, >> >> From a colocalization perspective.... >> >> Objective lenses are never perfect, not even really expensive ones that claim multi color chromatic correction. >> When it comes to imaging diffraction limited objects with a 1.4 NA lens on a widefield system >> there will still be a unique colour shift in all three spatial dimensions for that particular lens, >> and even how it is screwed in. >> >> Also the alignment of the fluorescence filters in different cubes might bend different channels to slightly different places on the detector. >> On the OMX you can even adjust this... normally its fixed on most systems but might also be tweakable. >> >> Olympus does all the chromatic correction in the lens, but others do some in the tube lens. >> This is why its not a good idea to screw a Zeiss c-apochromat 100x 1.6 oil into an olympus stand. >> >> So we are left with a systematic error that can be measured, actually to a precision of 10s of nm by Gaussian fitting bead images (like you do in PALM/SSTOR type super resolution and single molecule tracking. ) >> >> On a single point scanning confocal the matter is made worse by the more complicated optics. >> The Zeiss 510 has a different pinhole per channel... and depending on their positions and correct setup, >> the images of the bead in different channels is also affected by pinhole position settings. >> >> Nowadays with the Zeiss 7xx series and on the Olympus FV1000 and other modern point scanners there is only 1 pinhole >> for the emission light, so that problem is suppressed. >> BUT, the 405 lasers come in through a different fibre than the Vis lasers.. and so there is a collimation adjustment to make >> Same for 2photon lasers.... getting the near UV and IR laser spots to coincide exactly with the vis lines is hard... so its best to get is as close as is reasonable to do, then measure the residual error then correct for it. >> >> There always remain measurable errors (unless you were really really lucky and happen to have a perfect lens, which is a 1 in a 1000 chance at best, plus perfectly aligned fluorescence filters) >> >> OK, so at the simple level, we can assume that the whole field of view is shifted by some 3D shift per channel compared to the reference, usually green channel. So if we measure a few 1 or 0.5 or so micron beads (no need to use tiny beads here) near the centre of the field of view (where optics are best) we can calculate or guestimate the center of mass of the bead images in each channel, and work out the shift vectors needed. These will be between 10-1000 nm in xy and up to 500 or even 2000 micron in z. >> BUT, they will never be whole xy pixels or z slice shifts... forget that idea its too crude. >> >> see the 2 slides titled >> "Check with multi-colour beads" >> at http://ifn.mpi-cbg.de/wiki/ifn/images/0/0d/QuantitativeColocAnalysis-12-2010.pdf >> >> Now we can use Erik M's TransformJ Translate plugin in Fiji/ImageJ to do a sub pixel resolution shift for for each channel, as we have measured those shifts. >> http://www.imagescience.org/meijering/software/transformj/translate.html >> Use a nice interpolation method to avoid smashing the information in your images - eg quintic B-spline interpolation >> Or one could imaging a Fourier based method... (using a phase shift?) THe 3D stitching plugin contains stuff that might do this but its not exposed in the GUIs... so scripting would be required there. >> >> Moving the images using a whole pixel or z plane shift will not be precise enough for high res colocalization analysis. >> >> At a higher level there may also be a magnification difference between the different colour channels. >> On multi camera systems, like OMX or some HCS system there WILL be a rotation angle difference between the channel images. >> TransformJ Affine plugin might do the job, which can also expand or shrink to e image to accommodate different magnification and rotation. >> >> Even worse could be a non linear warp of the image that is different per channel. >> Again these are measurable and fixable with some effort, eg using bUnwarpJ >> >> So depending how precise you need to be, over how large a field of view... the difficulty of the correction varies. >> >> We could add this kind of colour shift correction to the new Coloc_2 plugin... by reusing TransformJ and maybe also bUnwarpJ. >> >> Its a whole world of fun. >> >> cheers >> >> Dan >> >> >> >> On Oct 20, 2011, at 6:02 AM, IMAGEJ automatic digest system wrote: >> >>> Date: Wed, 19 Oct 2011 09:44:25 -0400 >>> From: "Cammer, Michael" <[hidden email]> >>> Subject: Re: Plugin to correct for objective axial chromatic aberration? >>> >>> A few years ago we were using a CFP-YFP biosensor in fixed cells with a 60X N.A. 1.4 Olympus objective on an IX 70. We found that we were getting very bright edge artifacts in the ratio images. So I checked the system with 0.2 um beads in the same mounting media. I found that shifting the focus 0.3 um for CFP from the YFP image solved the problem. >>> >>> We had a similar problem last year imaging Alexa 488 and Cy5 with a DV2. The vendor assisted us and we found that by sandwiching a 1000 mm lens in with one of the filters, the focus was in registration. >>> >>> I don't see how we could have solved either of these focus issues post image collection. >>> >>> ________________________________________________________ >>> Michael Cammer, Assistant Research Scientist >>> Skirball Institute of Biomolecular Medicine >>> Lab: (212) 263-3208 Cell: (914) 309-3270 >>> >>> -----Original Message----- >>> From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of John Oreopoulos >>> Sent: Monday, October 17, 2011 9:26 PM >>> To: [hidden email] >>> Subject: Plugin to correct for objective axial chromatic aberration? >>> >>> My apologies as I suspect this question may have been posed to the listserver before, but I could not find an answer in the discussion archives... >>> >>> Suppose I have a well aligned microscope system and have made every effort to minimize refractive index mistmatch, etc. Further suppose then that I perform PSF measurements on some multi-colored sub-diffraction sized fluorescent microbeads in two or three different spectral channels. In this very unfortunate case, finally suppose that I measure an axial chromatic shift of the beads in the different spectral channels (even it may be an apochromat objective). >>> >>> My question is, once I know the amount of axial shift in each channel for this particular objective lens, is there a way to correct and realign images in the z-direction through image processing post-acqisution? Has anyone developed an ImageJ or FIJI plugin to do that? >>> >>> I would think that this is a very important aspect of colocalization (since colocalization is really a 3D problem, not just 2D), or other ratiometric fluorescence measurements such as FRET or calcium imaging. Surely the solution is not just to get a better chormatically corrected objective, is it? My sense is that this would be a problem that the experts in deconvolution fluorescence microscopy have already tackled as well. Is that the case? Is there any literature on this topic? >>> >>> Thanks for educating me if you can! >>> >>> John Oreopoulos >>> Research Assistant >>> Spectral Applied Research >>> Richmond Hill, Ontario >>> Canada >> >> Dr. Daniel James White BSc. (Hons.) PhD >> >> Leader - Image Processing Facility, >> Senior Microscopist, >> Light Microscopy Facility. >> >> Max Planck Institute of Molecular Cell Biology and Genetics >> Pfotenhauerstrasse 108 >> 01307 DRESDEN >> Germany >> >> +49 (0)15114966933 (German Mobile) >> +49 (0)351 210 2627 (Work phone at MPI-CBG) >> +49 (0)351 210 1078 (Fax MPI-CBG LMF) >> chalkie666 Skype >> http://www.bioimagexd.net BioImageXD >> http://fiji.sc Fiji - is just ImageJ (Batteries Included) >> http://www.chalkie.org.uk Dan's Homepages >> https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) >> dan (at) chalkie.org.uk >> ( white (at) mpi-cbg.de ) >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> > > Dr. Daniel James White BSc. (Hons.) PhD > > Leader - Image Processing Facility, > Senior Microscopist, > Light Microscopy Facility. > > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > chalkie666 Skype > http://www.bioimagexd.net BioImageXD > http://fiji.sc Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) > > > > > > > > > > > > > > > > > |
In reply to this post by Daniel James White
On Fri, Oct 21, 2011 at 16:37, Daniel James White <[hidden email]> wrote:
> Hi Folks, > > The author of the amazing SPIM registration / fusion plugins for multi > angle z stacks, Stephan Preibisch, > just informed me that the SPIM plugins will do the job of calculating the > 3D rigid or even affine transform between two or more z stacks of beads - > from eg different colour channels > and this transform can be appiled to other images easily.... so measure > with beads, then apply transform to real images. > > Would it be possible to have more detailled instructions on how to proceed for general channel registration with the SPIM plugins? I have a five-channel 2D tiff stack of a beads slide (no Z stack, just on plane for each channel) and I want to compute the XY rigid transform to register all channels. I found the SPIM package quite confusing (the dialogs are huge), I assume because it is powerful/flexible/adapted to its main task. If this can be really useful, would it be useful that I try to extract the channel-registration part of the SPIM plugins to make it simpler and more general? Christophe |
Hi Christope
On Oct 26, 2011, at 9:31 AM, Christophe Leterrier wrote: > On Fri, Oct 21, 2011 at 16:37, Daniel James White <[hidden email]> wrote: > Hi Folks, > > The author of the amazing SPIM registration / fusion plugins for multi angle z stacks, Stephan Preibisch, > just informed me that the SPIM plugins will do the job of calculating the > 3D rigid or even affine transform between two or more z stacks of beads - from eg different colour channels > and this transform can be appiled to other images easily.... so measure with beads, then apply transform to real images. > > > Hi Dan (and Stephan) > > Would it be possible to have more detailled instructions on how to proceed for general channel registration with the SPIM plugins? I have a five-channel 2D tiff stack of a beads slide (no Z stack, just on plane for each channel) and I want to compute the XY rigid transform to register all channels. > > I found the SPIM package quite confusing (the dialogs are huge), I assume because it is powerful/flexible/adapted to its main task. If this can be really useful, would it be useful that I try to extract the channel-registration part of the SPIM plugins to make it simpler and more general? > > Christophe Once I figure it out, i will spread the word! D Dr. Daniel James White BSc. (Hons.) PhD Leader - Image Processing Facility, Senior Microscopist, Light Microscopy Facility. Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) chalkie666 Skype http://www.bioimagexd.net BioImageXD http://fiji.sc Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Biopolis Dresden Imaging Platform (BioDIP) dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
In reply to this post by lechristophe
Hi Christophe,
we have already done this once for aligning two camera images using a perspective transformation based on dust particles in the images. I hope that StephanP can find the respective script back. Starting from that, we can easily make a small special purpose plugin for point-cloud based pairwise 2d image alignment. This plugin can then benefit from all the transformation models available in Fiji that are: affine and subclasses (translation, rigid body, similarity), homography, n-th order polynomial, moving least squares warp If ImageJ only had double precision PointRois (how difficult would it be to add that actually?), we could make it even more flexible and extract only the point correspondences similar to the plugins in http://fiji.sc/wiki/index.php/Feature_Extraction Those correspondences can then be used to warp an image using the plugin http://fiji.sc/Landmark_Correspondences Currently, this plugin suffers from integer precision PointRoi correspondences and, therefore, cannot be used for a high precision application as yours. Best, StephanS On Wed, 2011-10-26 at 09:31 +0200, Christophe Leterrier wrote: > On Fri, Oct 21, 2011 at 16:37, Daniel James White <[hidden email]> wrote: > > > Hi Folks, > > > > The author of the amazing SPIM registration / fusion plugins for multi > > angle z stacks, Stephan Preibisch, > > just informed me that the SPIM plugins will do the job of calculating the > > 3D rigid or even affine transform between two or more z stacks of beads - > > from eg different colour channels > > and this transform can be appiled to other images easily.... so measure > > with beads, then apply transform to real images. > > > > > Hi Dan (and Stephan) > > Would it be possible to have more detailled instructions on how to proceed > for general channel registration with the SPIM plugins? I have a > five-channel 2D tiff stack of a beads slide (no Z stack, just on plane for > each channel) and I want to compute the XY rigid transform to register all > channels. > > I found the SPIM package quite confusing (the dialogs are huge), I assume > because it is powerful/flexible/adapted to its main task. If this can be > really useful, would it be useful that I try to extract the > channel-registration part of the SPIM plugins to make it simpler and more > general? > > Christophe |
Hi,
of course I still have it :) Sorry for not replying so fast, I just moved this week to my new post-doc place which is a bit distracting ... I will try to write this code as a plugin this week, it should be not too difficult. We could internally extend RointRois to double precision as Stephan and me once discussed and then simply try to cast in the Landmark_correspondence plugin. If coordinates were manually changed meanwhile I would simply discard the double precision. The method could check if the rounded double positions still reflects the integer coordinates. This is quite an ugly hack, but still maybe better then implementing the fusion twice I'd say... Let me know what you think, StephanP (aka Steffi) -------------------------------------------------- From: "Stephan Saalfeld" <[hidden email]> Sent: Wednesday, October 26, 2011 4:21 AM To: <[hidden email]> Subject: Re: Plugin to correct for objective axial chromatic aberration? > Hi Christophe, > > we have already done this once for aligning two camera images using a > perspective transformation based on dust particles in the images. I > hope that StephanP can find the respective script back. Starting from > that, we can easily make a small special purpose plugin for point-cloud > based pairwise 2d image alignment. This plugin can then benefit from > all the transformation models available in Fiji that are: > > affine and subclasses (translation, rigid body, similarity), homography, > n-th order polynomial, moving least squares warp > > If ImageJ only had double precision PointRois (how difficult would it be > to add that actually?), we could make it even more flexible and extract > only the point correspondences similar to the plugins in > > http://fiji.sc/wiki/index.php/Feature_Extraction > > Those correspondences can then be used to warp an image using the plugin > > http://fiji.sc/Landmark_Correspondences > > Currently, this plugin suffers from integer precision PointRoi > correspondences and, therefore, cannot be used for a high precision > application as yours. > > Best, > StephanS > > > > > On Wed, 2011-10-26 at 09:31 +0200, Christophe Leterrier wrote: >> On Fri, Oct 21, 2011 at 16:37, Daniel James White <[hidden email]> >> wrote: >> >> > Hi Folks, >> > >> > The author of the amazing SPIM registration / fusion plugins for multi >> > angle z stacks, Stephan Preibisch, >> > just informed me that the SPIM plugins will do the job of calculating >> > the >> > 3D rigid or even affine transform between two or more z stacks of >> > beads - >> > from eg different colour channels >> > and this transform can be appiled to other images easily.... so measure >> > with beads, then apply transform to real images. >> > >> > >> Hi Dan (and Stephan) >> >> Would it be possible to have more detailled instructions on how to >> proceed >> for general channel registration with the SPIM plugins? I have a >> five-channel 2D tiff stack of a beads slide (no Z stack, just on plane >> for >> each channel) and I want to compute the XY rigid transform to register >> all >> channels. >> >> I found the SPIM package quite confusing (the dialogs are huge), I assume >> because it is powerful/flexible/adapted to its main task. If this can be >> really useful, would it be useful that I try to extract the >> channel-registration part of the SPIM plugins to make it simpler and more >> general? >> >> Christophe > |
In reply to this post by Stephan Saalfeld
On Oct 26, 2011, at 4:21 AM, Stephan Saalfeld wrote:
> Hi Christophe, > > we have already done this once for aligning two camera images using a > perspective transformation based on dust particles in the images. I > hope that StephanP can find the respective script back. Starting from > that, we can easily make a small special purpose plugin for point-cloud > based pairwise 2d image alignment. This plugin can then benefit from > all the transformation models available in Fiji that are: > > affine and subclasses (translation, rigid body, similarity), homography, > n-th order polynomial, moving least squares warp > > If ImageJ only had double precision PointRois (how difficult would it be > to add that actually?), .... ImageJ 1.46, due in a few weeks, is going to have sub-pixel resolution PointRois and PolygonRois. -wayne |
Hi Wayne,
that's great news. Just one comment before things are written to stone. Could those double precision Rois be parent classes (or may be an implemented interface) of the existing integer precision Rois. That way, one could always ask for double precision coordinates which isn't actually a problem for the integer versions. Pseudo UML: class inheritance (more interference) DoublePointRoi <- PointRoi interface DoublePoints <- PointRoi ^ | DoublePointRoi Thanks for your rapid help and best regards, StephanS On Wed, 2011-10-26 at 09:12 -0400, Rasband, Wayne (NIH/NIMH) [E] wrote: > On Oct 26, 2011, at 4:21 AM, Stephan Saalfeld wrote: > > > Hi Christophe, > > > > we have already done this once for aligning two camera images using a > > perspective transformation based on dust particles in the images. I > > hope that StephanP can find the respective script back. Starting from > > that, we can easily make a small special purpose plugin for point-cloud > > based pairwise 2d image alignment. This plugin can then benefit from > > all the transformation models available in Fiji that are: > > > > affine and subclasses (translation, rigid body, similarity), homography, > > n-th order polynomial, moving least squares warp > > > > If ImageJ only had double precision PointRois (how difficult would it be > > to add that actually?), .... > > ImageJ 1.46, due in a few weeks, is going to have sub-pixel resolution PointRois and PolygonRois. > > -wayne |
On Oct 26, 2011, at 2:09 PM, Stephan Saalfeld wrote:
> Hi Wayne, > > that's great news. Just one comment before things are written to stone. > Could those double precision Rois be parent classes (or may be an > implemented interface) of the existing integer precision Rois. That > way, one could always ask for double precision coordinates which isn't > actually a problem for the integer versions. There are not going to be any new classes. The coordinates will be single precision and you will retrieve them using the getFloatPolygon() method. This method already exists and it works with most selection types. You can write plugins now that use it and in a few weeks the coordinates will have sub-pixel resolution. Spline fit selections already have sub-pixel resolution. Here is some JavaScript code that shows how to use it: imp = IJ.getImage(); roi = imp.getRoi(); if (roi==null) IJ.error("No selection"); p = roi.getFloatPolygon(); for (i=0; i<p.npoints; i++) print(i+" "+p.xpoints[i]+" "+p.ypoints[i]); -wayne > Pseudo UML: > > class inheritance (more interference) > > DoublePointRoi <- PointRoi > > interface > > DoublePoints <- PointRoi > ^ > | > DoublePointRoi > > Thanks for your rapid help and best regards, > StephanS > > > > > On Wed, 2011-10-26 at 09:12 -0400, Rasband, Wayne (NIH/NIMH) [E] wrote: >> On Oct 26, 2011, at 4:21 AM, Stephan Saalfeld wrote: >> >>> Hi Christophe, >>> >>> we have already done this once for aligning two camera images using a >>> perspective transformation based on dust particles in the images. I >>> hope that StephanP can find the respective script back. Starting from >>> that, we can easily make a small special purpose plugin for point-cloud >>> based pairwise 2d image alignment. This plugin can then benefit from >>> all the transformation models available in Fiji that are: >>> >>> affine and subclasses (translation, rigid body, similarity), homography, >>> n-th order polynomial, moving least squares warp >>> >>> If ImageJ only had double precision PointRois (how difficult would it be >>> to add that actually?), .... >> >> ImageJ 1.46, due in a few weeks, is going to have sub-pixel resolution PointRois and PolygonRois. >> >> -wayne |
Perfect---thanks a lot! Looking forward to it.
I assume that there will be a constructor PointRoi(float[] ox, float[] oy, int points) ? Best, Stephan On Wed, 2011-10-26 at 16:06 -0400, Rasband, Wayne (NIH/NIMH) [E] wrote: > On Oct 26, 2011, at 2:09 PM, Stephan Saalfeld wrote: > > > Hi Wayne, > > > > that's great news. Just one comment before things are written to stone. > > Could those double precision Rois be parent classes (or may be an > > implemented interface) of the existing integer precision Rois. That > > way, one could always ask for double precision coordinates which isn't > > actually a problem for the integer versions. > > There are not going to be any new classes. The coordinates will be > single precision and you will retrieve them using the > getFloatPolygon() method. This method already exists and it works with > most selection types. You can write plugins now that use it and in a > few weeks the coordinates will have sub-pixel resolution. Spline fit > selections already have sub-pixel resolution. Here is some JavaScript > code that shows how to use it: > > imp = IJ.getImage(); > roi = imp.getRoi(); > if (roi==null) > IJ.error("No selection"); > p = roi.getFloatPolygon(); > for (i=0; i<p.npoints; i++) > print(i+" "+p.xpoints[i]+" "+p.ypoints[i]); > > -wayne > > > > Pseudo UML: > > > > class inheritance (more interference) > > > > DoublePointRoi <- PointRoi > > > > interface > > > > DoublePoints <- PointRoi > > ^ > > | > > DoublePointRoi > > > > Thanks for your rapid help and best regards, > > StephanS > > > > > > > > > > On Wed, 2011-10-26 at 09:12 -0400, Rasband, Wayne (NIH/NIMH) [E] wrote: > >> On Oct 26, 2011, at 4:21 AM, Stephan Saalfeld wrote: > >> > >>> Hi Christophe, > >>> > >>> we have already done this once for aligning two camera images using a > >>> perspective transformation based on dust particles in the images. I > >>> hope that StephanP can find the respective script back. Starting from > >>> that, we can easily make a small special purpose plugin for point-cloud > >>> based pairwise 2d image alignment. This plugin can then benefit from > >>> all the transformation models available in Fiji that are: > >>> > >>> affine and subclasses (translation, rigid body, similarity), homography, > >>> n-th order polynomial, moving least squares warp > >>> > >>> If ImageJ only had double precision PointRois (how difficult would it be > >>> to add that actually?), .... > >> > >> ImageJ 1.46, due in a few weeks, is going to have sub-pixel resolution PointRois and PolygonRois. > >> > >> -wayne |
Free forum by Nabble | Edit this page |