Segmentation of DIC or Hoffman Modulation Contrast images - help please
Locked
Segmentation of DIC or Hoffman Modulation Contrast images - help please
 
|
Hi All,
I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). Look forward to hearing any suggestions! Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
HMC_test.jpg (141K)
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Try converting to 8-bit and running a variance filter
(Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. --aryeh On 5/1/15 7:37 AM, Jacqui Ross wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. > > I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- Aryeh Weiss Faculty of Engineering Bar Ilan University Ramat Gan 52900 Israel Ph: 972-3-5317638 FAX: 972-3-7384051 -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
FW: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hi All,
Just realised my reply to Aryeh didn't get to the listserv. Please see below. Cheers, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: Jacqui Ross Sent: Friday, 1 May 2015 5:48 p.m. To: '[hidden email]' Subject: RE: Segmentation of DIC or Hoffman Modulation Contrast images - help please Hi Aryeh, Thanks for your reply. I did try using a variance filter (the built-in one under Process - Filters - Variance) with different radii but I was unable to achieve a good result. The resultant circles were often incomplete so that when I then converted to binary, I had to do a lot of additional processing (Closing, filling holes, etc.) and then the outlines weren't very accurate. I know that you presented on the Trainable Weka Segmentation at the ImageJ conference that I attended a couple of years ago. My notes on that weren't fantastic (I pulled them out!) but in your case you also had some fluorescence labelling to help inform the segmentation. Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: Aryeh Weiss [mailto:[hidden email]] On Behalf Of Aryeh Weiss Sent: Friday, 1 May 2015 5:07 p.m. To: Jacqui Ross Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Try converting to 8-bit and running a variance filter (Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. --aryeh On 5/1/15 7:37 AM, Jacqui Ross wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. > > I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- Aryeh Weiss Faculty of Engineering Bar Ilan University Ramat Gan 52900 Israel Ph: 972-3-5317638 FAX: 972-3-7384051 -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
On 5/1/15 8:48 AM, Jacqui Ross wrote:
> Hi Aryeh, > > Thanks for your reply. I did try using a variance filter (the built-in one under Process - Filters - Variance) with different radii but I was unable to achieve a good result. The resultant circles were often incomplete so that when I then converted to binary, I had to do a lot of additional processing (Closing, filling holes, etc.) and then the outlines weren't very accurate. Yes -- these methods are better at marking objects than getting accurate boundaries. You might be able to use the inaccurate segmentation that produces as a mask against the original variance image, which produces reasonable arcs around your in-focus cells. > I know that you presented on the Trainable Weka Segmentation at the ImageJ conference that I attended a couple of years ago. My notes on that weren't fantastic (I pulled them out!) but in your case you also had some fluorescence labelling to help inform the segmentation. That problem was easier than yours (isn't it always that way?). The texture was much better defined, and I did not have a continuum of out-of-focus or partially out-of-focus cells. There was a DAPI channel which I used to exclude artitfacts which were not cells (since they did not contain nuclei). However, that will not help you, since your out of focus cells are still cells. You might have an easier time here if you can acquire a z-stack (even though it is wide-field) and use the EDF plugin or similar to have all of your cells in-focus. That would be easier to segment. Alternatively, if you choose to be very strict in your segmentation (ie, take only the cells that are really sharp and textured), then I think you will be able to variance and it relatives to get an accurate segmentation and boundary. Best regards --aryeh > > -----Original Message----- > From: Aryeh Weiss [mailto:[hidden email]] On Behalf Of Aryeh Weiss > Sent: Friday, 1 May 2015 5:07 p.m. > To: Jacqui Ross > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please > > > Try converting to 8-bit and running a variance filter > (Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. > > --aryeh > > On 5/1/15 7:37 AM, Jacqui Ross wrote: >> Hi All, >> >> I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. >> >> In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. >> These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. >> >> I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. >> >> In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. >> >> The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. >> >> I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). >> >> Look forward to hearing any suggestions! >> >> Kind regards, >> >> Jacqui >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -- >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html > > -- > Aryeh Weiss > Faculty of Engineering > Bar Ilan University > Ramat Gan 52900 Israel > > Ph: 972-3-5317638 > FAX: 972-3-7384051 > > > . > -- Aryeh Weiss Faculty of Engineering Bar Ilan University Ramat Gan 52900 Israel Ph: 972-3-5317638 FAX: 972-3-7384051 -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
I would consider an approach that is one level higher than raw image processing. DIC images produce distinctive doublets (a bright region adjacent to a darker region - always (for the same setup) at the same angle.
Once you can identify a cell from it’s doublet, and have a reasonable guess as to its size, finding the boundary and estimating area should become easier. Just a thought. My first try might be to design a custom convolution kernel that responds preferentially to doublets at a fixed angle. There is still the problem of estimating the angle (if it’s not stable) but that should be doable based either on meta-information or user input. Are all the cells as well-behaved as these? I second the idea of using a texture measurement to eliminate regions of the image that are out of focus (variance is one - I would perhaps go a step further and use coefficient of variation - stdDev/mean - but there are others) -- Kenneth Sloan [hidden email] Vision is the art of seeing what is invisible to others. > On May 1, 2015, at 08:02 , Aryeh Weiss <[hidden email]> wrote: > > On 5/1/15 8:48 AM, Jacqui Ross wrote: >> Hi Aryeh, >> >> Thanks for your reply. I did try using a variance filter (the built-in one under Process - Filters - Variance) with different radii but I was unable to achieve a good result. The resultant circles were often incomplete so that when I then converted to binary, I had to do a lot of additional processing (Closing, filling holes, etc.) and then the outlines weren't very accurate. > Yes -- these methods are better at marking objects than getting accurate boundaries. You might be able to use the inaccurate segmentation that produces as a mask against the original variance image, which produces reasonable arcs around your in-focus cells. >> I know that you presented on the Trainable Weka Segmentation at the ImageJ conference that I attended a couple of years ago. My notes on that weren't fantastic (I pulled them out!) but in your case you also had some fluorescence labelling to help inform the segmentation. > That problem was easier than yours (isn't it always that way?). The texture was much better defined, and I did not have a continuum of out-of-focus or partially out-of-focus cells. There was a DAPI channel which I used to exclude artitfacts which were not cells (since they did not contain nuclei). However, that will not help you, since your out of focus cells are still cells. > > You might have an easier time here if you can acquire a z-stack (even though it is wide-field) and use the EDF plugin or similar to have all of your cells in-focus. That would be easier to segment. Alternatively, if you choose to be very strict in your segmentation (ie, take only the cells that are really sharp and textured), then I think you will be able to variance and it relatives to get an accurate segmentation and boundary. > > Best regards > --aryeh > >> >> -----Original Message----- >> From: Aryeh Weiss [mailto:[hidden email]] On Behalf Of Aryeh Weiss >> Sent: Friday, 1 May 2015 5:07 p.m. >> To: Jacqui Ross >> Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please >> >> >> Try converting to 8-bit and running a variance filter >> (Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. >> >> --aryeh >> >> On 5/1/15 7:37 AM, Jacqui Ross wrote: >>> Hi All, >>> >>> I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. >>> >>> In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. >>> These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. >>> >>> I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. >>> >>> In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. >>> >>> The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. >>> >>> I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). >>> >>> Look forward to hearing any suggestions! >>> >>> Kind regards, >>> >>> Jacqui >>> Jacqueline Ross >>> Biomedical Imaging Microscopist >>> Biomedical Imaging Research Unit >>> School of Medical Sciences >>> Faculty of Medical & Health Sciences >>> The University of Auckland >>> Private Bag 92019 >>> Auckland 1142, NEW ZEALAND >>> >>> Tel: 64 9 923 7438 >>> Fax: 64 9 373 7484 >>> >>> http://www.fmhs.auckland.ac.nz/sms/biru/ >>> >>> >>> -- >>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html >> >> -- >> Aryeh Weiss >> Faculty of Engineering >> Bar Ilan University >> Ramat Gan 52900 Israel >> >> Ph: 972-3-5317638 >> FAX: 972-3-7384051 >> >> >> . >> > > > -- > Aryeh Weiss > Faculty of Engineering > Bar Ilan University > Ramat Gan 52900 Israel > > Ph: 972-3-5317638 > FAX: 972-3-7384051 > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hi Jacqui,
These images are particularly difficult to segment because the edges are assymetric --dark on one side, and light on the other. I was able to get some enhancement by using an unsharp mask (on your image, pixel radius of 100, weight .60), followed by the "find edges" convolution. It wasn't perfect, but might help. Joel Joel B. Sheffield, Ph.D Department of Biology Temple University Philadelphia, PA 19122 Voice: 215 204 8839 e-mail: [hidden email] URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* On Fri, May 1, 2015 at 12:37 AM, Jacqui Ross <[hidden email]> wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast > images of cells. She's primarily interested in changes in diameter. The > cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus > others which are out of focus which we don't want to measure as any > measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, > already knows. I've tried lots of different filters (edge, etc.) , FFT > filtering and the Trainable Weka Segmentation but have been unable to > achieve good enough results to be able to then threshold the cells > automatically. > > I've come to the end of the line for now so am asking for your expert help > in case anyone has some suggestions:). I note that in 2006 Monique Vasseur > offered some DIC images to a PhD student called Daniel Mauch in Germany but > I'm not sure if anything came of that project. There are a few papers out > there (some mention Hilbert Transform, FFT) but I haven't been successful > in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction > plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small > radius (5) to flatten the background and out of focus cells while > preserving the in focus cell outlines. We can then use the Cell Magic Wand > (Thanks Theo!) to create selections that can be loaded into the ROI Manager > and then measured. This works really well but requires that the cells be > selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also > split the channels from the colour image if needed and use the channel > image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been > cropped out of a larger image so that it's not too big and it's pink > because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html > -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Again - I’ll promote taking a higher level approach. Simple SEGMENTATION should be able to pull out “Bright blobs” and “Dark blobs” - segmenting the CELLS requires combining that evidence (and then perhaps going back to the image data in “verification vision” mode - making predictions about what the cell boundary will look like, and localizing it. But, this is probably beyond the scope of a script combining standard image processing operators.
-- Kenneth Sloan [hidden email] Vision is the art of seeing what is invisible to others. > On May 1, 2015, at 09:10 , JOEL B. SHEFFIELD <[hidden email]> wrote: > > Hi Jacqui, > > These images are particularly difficult to segment because the edges are > assymetric --dark on one side, and light on the other. I was able to get > some enhancement by using an unsharp mask (on your image, pixel radius of > 100, weight .60), followed by the "find edges" convolution. It wasn't > perfect, but might help. > > Joel > > > > Joel B. Sheffield, Ph.D > Department of Biology > Temple University > Philadelphia, PA 19122 > Voice: 215 204 8839 > e-mail: [hidden email] > URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* > > On Fri, May 1, 2015 at 12:37 AM, Jacqui Ross <[hidden email]> > wrote: > >> Hi All, >> >> I'm helping a PhD student with analysing some Hoffman modulation contrast >> images of cells. She's primarily interested in changes in diameter. The >> cells are embedded in a 3D matrix and compression is being applied. >> >> In the images, there are nice cells in focus with clear boundaries, plus >> others which are out of focus which we don't want to measure as any >> measurements won't be accurate. >> These kind of images are really tricky to segment as anyone who has tried, >> already knows. I've tried lots of different filters (edge, etc.) , FFT >> filtering and the Trainable Weka Segmentation but have been unable to >> achieve good enough results to be able to then threshold the cells >> automatically. >> >> I've come to the end of the line for now so am asking for your expert help >> in case anyone has some suggestions:). I note that in 2006 Monique Vasseur >> offered some DIC images to a PhD student called Daniel Mauch in Germany but >> I'm not sure if anything came of that project. There are a few papers out >> there (some mention Hilbert Transform, FFT) but I haven't been successful >> in implementing anything from those papers as yet. >> >> In the meantime, my solution is to use the Pseudo flat field correction >> plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small >> radius (5) to flatten the background and out of focus cells while >> preserving the in focus cell outlines. We can then use the Cell Magic Wand >> (Thanks Theo!) to create selections that can be loaded into the ROI Manager >> and then measured. This works really well but requires that the cells be >> selected manually. >> >> The Cell Magic Wand Tool works on the colour or grayscale so we can also >> split the channels from the colour image if needed and use the channel >> image with the most contrast. >> >> I've attached an image in case anyone has any ideas. The image has been >> cropped out of a larger image so that it's not too big and it's pink >> because there's cell culture medium there (in case anyone was wondering..). >> >> Look forward to hearing any suggestions! >> >> Kind regards, >> >> Jacqui >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -- >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html >> > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Aryeh Weiss
Hi Aryeh,
Thanks for your suggestions. Ralph also suggested acquiring a z stack. The problem is that we are applying compression from the side to a 3D deformable matrix in which the cells are embedded. This means that the cells move sideways so a z stack would then require XY translation/registration to get the same cells. We could do it in that it's static compression so could hopefully work out which cells were the same ones if we had some fiducial markers (we don't!). If I had been involved in designing the initial experiments, I would have suggested compressing from the top, which would have resulted in less XY translation but the same deformation and we could have then used the z stack idea. Kind regards, Jacqui ________________________________________ From: Aryeh Weiss [[hidden email]] on behalf of Aryeh Weiss [[hidden email]] Sent: 02 May 2015 01:02 To: Jacqui Ross; ImageJ Interest Group Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please On 5/1/15 8:48 AM, Jacqui Ross wrote: > Hi Aryeh, > > Thanks for your reply. I did try using a variance filter (the built-in one under Process - Filters - Variance) with different radii but I was unable to achieve a good result. The resultant circles were often incomplete so that when I then converted to binary, I had to do a lot of additional processing (Closing, filling holes, etc.) and then the outlines weren't very accurate. Yes -- these methods are better at marking objects than getting accurate boundaries. You might be able to use the inaccurate segmentation that produces as a mask against the original variance image, which produces reasonable arcs around your in-focus cells. > I know that you presented on the Trainable Weka Segmentation at the ImageJ conference that I attended a couple of years ago. My notes on that weren't fantastic (I pulled them out!) but in your case you also had some fluorescence labelling to help inform the segmentation. That problem was easier than yours (isn't it always that way?). The texture was much better defined, and I did not have a continuum of out-of-focus or partially out-of-focus cells. There was a DAPI channel which I used to exclude artitfacts which were not cells (since they did not contain nuclei). However, that will not help you, since your out of focus cells are still cells. You might have an easier time here if you can acquire a z-stack (even though it is wide-field) and use the EDF plugin or similar to have all of your cells in-focus. That would be easier to segment. Alternatively, if you choose to be very strict in your segmentation (ie, take only the cells that are really sharp and textured), then I think you will be able to variance and it relatives to get an accurate segmentation and boundary. Best regards --aryeh > > -----Original Message----- > From: Aryeh Weiss [mailto:[hidden email]] On Behalf Of Aryeh Weiss > Sent: Friday, 1 May 2015 5:07 p.m. > To: Jacqui Ross > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please > > > Try converting to 8-bit and running a variance filter > (Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. > > --aryeh > > On 5/1/15 7:37 AM, Jacqui Ross wrote: >> Hi All, >> >> I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. >> >> In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. >> These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. >> >> I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. >> >> In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. >> >> The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. >> >> I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). >> >> Look forward to hearing any suggestions! >> >> Kind regards, >> >> Jacqui >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -- >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html > > -- > Aryeh Weiss > Faculty of Engineering > Bar Ilan University > Ramat Gan 52900 Israel > > Ph: 972-3-5317638 > FAX: 972-3-7384051 > > > . > -- Aryeh Weiss Faculty of Engineering Bar Ilan University Ramat Gan 52900 Israel Ph: 972-3-5317638 FAX: 972-3-7384051 -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Kenneth R Sloan
Hi Kenneth,
Thanks for your suggestions. The cells are primarily almost circular. You are correct in your comments about the edges. It's easy to segment out the bright or dark crescent shapes but then the circles are incomplete. I'm not sure what you mean by a higher level approach - something like machine learning? I'll have to keep persevering. Kind regards, Jacqui ________________________________________ From: ImageJ Interest Group [[hidden email]] on behalf of Kenneth R Sloan [[hidden email]] Sent: 02 May 2015 02:36 To: [hidden email] Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Again - I’ll promote taking a higher level approach. Simple SEGMENTATION should be able to pull out “Bright blobs” and “Dark blobs” - segmenting the CELLS requires combining that evidence (and then perhaps going back to the image data in “verification vision” mode - making predictions about what the cell boundary will look like, and localizing it. But, this is probably beyond the scope of a script combining standard image processing operators. -- Kenneth Sloan [hidden email] Vision is the art of seeing what is invisible to others. > On May 1, 2015, at 09:10 , JOEL B. SHEFFIELD <[hidden email]> wrote: > > Hi Jacqui, > > These images are particularly difficult to segment because the edges are > assymetric --dark on one side, and light on the other. I was able to get > some enhancement by using an unsharp mask (on your image, pixel radius of > 100, weight .60), followed by the "find edges" convolution. It wasn't > perfect, but might help. > > Joel > > > > Joel B. Sheffield, Ph.D > Department of Biology > Temple University > Philadelphia, PA 19122 > Voice: 215 204 8839 > e-mail: [hidden email] > URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* > > On Fri, May 1, 2015 at 12:37 AM, Jacqui Ross <[hidden email]> > wrote: > >> Hi All, >> >> I'm helping a PhD student with analysing some Hoffman modulation contrast >> images of cells. She's primarily interested in changes in diameter. The >> cells are embedded in a 3D matrix and compression is being applied. >> >> In the images, there are nice cells in focus with clear boundaries, plus >> others which are out of focus which we don't want to measure as any >> measurements won't be accurate. >> These kind of images are really tricky to segment as anyone who has tried, >> already knows. I've tried lots of different filters (edge, etc.) , FFT >> filtering and the Trainable Weka Segmentation but have been unable to >> achieve good enough results to be able to then threshold the cells >> automatically. >> >> I've come to the end of the line for now so am asking for your expert help >> in case anyone has some suggestions:). I note that in 2006 Monique Vasseur >> offered some DIC images to a PhD student called Daniel Mauch in Germany but >> I'm not sure if anything came of that project. There are a few papers out >> there (some mention Hilbert Transform, FFT) but I haven't been successful >> in implementing anything from those papers as yet. >> >> In the meantime, my solution is to use the Pseudo flat field correction >> plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small >> radius (5) to flatten the background and out of focus cells while >> preserving the in focus cell outlines. We can then use the Cell Magic Wand >> (Thanks Theo!) to create selections that can be loaded into the ROI Manager >> and then measured. This works really well but requires that the cells be >> selected manually. >> >> The Cell Magic Wand Tool works on the colour or grayscale so we can also >> split the channels from the colour image if needed and use the channel >> image with the most contrast. >> >> I've attached an image in case anyone has any ideas. The image has been >> cropped out of a larger image so that it's not too big and it's pink >> because there's cell culture medium there (in case anyone was wondering..). >> >> Look forward to hearing any suggestions! >> >> Kind regards, >> >> Jacqui >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -- >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html >> > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Joel Sheffield
Hi Joel,
Yes, I've had similar "almost" success although I was running the UnSharp Mask after the background correction. Sometimes changing the order solves the problem. I'll try your suggestions. The difficulty I've had so far is that the texture inside the cells gets enhanced with the UnSharp Mask then creating boundaries inside the cells and creating half cells due also to the light/dark crescents. Kind regards, Jacqui ________________________________________ From: ImageJ Interest Group [[hidden email]] on behalf of JOEL B. SHEFFIELD [[hidden email]] Sent: 02 May 2015 02:10 To: [hidden email] Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Hi Jacqui, These images are particularly difficult to segment because the edges are assymetric --dark on one side, and light on the other. I was able to get some enhancement by using an unsharp mask (on your image, pixel radius of 100, weight .60), followed by the "find edges" convolution. It wasn't perfect, but might help. Joel Joel B. Sheffield, Ph.D Department of Biology Temple University Philadelphia, PA 19122 Voice: 215 204 8839 e-mail: [hidden email] URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* On Fri, May 1, 2015 at 12:37 AM, Jacqui Ross <[hidden email]> wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast > images of cells. She's primarily interested in changes in diameter. The > cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus > others which are out of focus which we don't want to measure as any > measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, > already knows. I've tried lots of different filters (edge, etc.) , FFT > filtering and the Trainable Weka Segmentation but have been unable to > achieve good enough results to be able to then threshold the cells > automatically. > > I've come to the end of the line for now so am asking for your expert help > in case anyone has some suggestions:). I note that in 2006 Monique Vasseur > offered some DIC images to a PhD student called Daniel Mauch in Germany but > I'm not sure if anything came of that project. There are a few papers out > there (some mention Hilbert Transform, FFT) but I haven't been successful > in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction > plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small > radius (5) to flatten the background and out of focus cells while > preserving the in focus cell outlines. We can then use the Cell Magic Wand > (Thanks Theo!) to create selections that can be loaded into the ROI Manager > and then measured. This works really well but requires that the cells be > selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also > split the channels from the colour image if needed and use the channel > image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been > cropped out of a larger image so that it's not too big and it's pink > because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html > -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Much simpler than “machine learning” - but it probably involves writing a small amount of Java code to build a custom plugin.
a) use standard operations to find a collection of bright “half-cells” b) use standard operations to find a collection of dark “half-cells” c) match up half-cells from a) and b) to create a collection of cells d) fit a disc to each cell (a center and a radius should be sufficient) e) for extra credit - go back to the image, predict whether an edge *should* be present and what sign it has (dark interior or bright interior) and find tune to actual visible boundary of the cell (at, say, 8 directions) As I type this, I realize that there’s a neat trick: you might consider the following pure image-processing algorithm: a) determine the correct value for middle gray b) calculate an “absolute value” image, by flipping all pixels darker than middle gray symetrically around middle gray c) scale intensities so that middle gray is mapped to black (leave white where it is) You now have an image of bright cells on a dark background. Find them. You will still have incomplete boundaries (contrast goes to zero at two places around the perimeter. That’s all I’ve got. Good luck. -- Kenneth Sloan [hidden email] Vision is the art of seeing what is invisible to others. > On May 2, 2015, at 19:43 , Jacqui Ross <[hidden email]> wrote: > > Hi Joel, > > Yes, I've had similar "almost" success although I was running the UnSharp Mask after the background correction. Sometimes changing the order solves the problem. > > I'll try your suggestions. The difficulty I've had so far is that the texture inside the cells gets enhanced with the UnSharp Mask then creating boundaries inside the cells and creating half cells due also to the light/dark crescents. > > Kind regards, > > Jacqui > ________________________________________ > From: ImageJ Interest Group [[hidden email]] on behalf of JOEL B. SHEFFIELD [[hidden email]] > Sent: 02 May 2015 02:10 > To: [hidden email] > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please > > Hi Jacqui, > > These images are particularly difficult to segment because the edges are > assymetric --dark on one side, and light on the other. I was able to get > some enhancement by using an unsharp mask (on your image, pixel radius of > 100, weight .60), followed by the "find edges" convolution. It wasn't > perfect, but might help. > > Joel > > > > Joel B. Sheffield, Ph.D > Department of Biology > Temple University > Philadelphia, PA 19122 > Voice: 215 204 8839 > e-mail: [hidden email] > URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* > > On Fri, May 1, 2015 at 12:37 AM, Jacqui Ross <[hidden email]> > wrote: > >> Hi All, >> >> I'm helping a PhD student with analysing some Hoffman modulation contrast >> images of cells. She's primarily interested in changes in diameter. The >> cells are embedded in a 3D matrix and compression is being applied. >> >> In the images, there are nice cells in focus with clear boundaries, plus >> others which are out of focus which we don't want to measure as any >> measurements won't be accurate. >> These kind of images are really tricky to segment as anyone who has tried, >> already knows. I've tried lots of different filters (edge, etc.) , FFT >> filtering and the Trainable Weka Segmentation but have been unable to >> achieve good enough results to be able to then threshold the cells >> automatically. >> >> I've come to the end of the line for now so am asking for your expert help >> in case anyone has some suggestions:). I note that in 2006 Monique Vasseur >> offered some DIC images to a PhD student called Daniel Mauch in Germany but >> I'm not sure if anything came of that project. There are a few papers out >> there (some mention Hilbert Transform, FFT) but I haven't been successful >> in implementing anything from those papers as yet. >> >> In the meantime, my solution is to use the Pseudo flat field correction >> plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small >> radius (5) to flatten the background and out of focus cells while >> preserving the in focus cell outlines. We can then use the Cell Magic Wand >> (Thanks Theo!) to create selections that can be loaded into the ROI Manager >> and then measured. This works really well but requires that the cells be >> selected manually. >> >> The Cell Magic Wand Tool works on the colour or grayscale so we can also >> split the channels from the colour image if needed and use the channel >> image with the most contrast. >> >> I've attached an image in case anyone has any ideas. The image has been >> cropped out of a larger image so that it's not too big and it's pink >> because there's cell culture medium there (in case anyone was wondering..). >> >> Look forward to hearing any suggestions! >> >> Kind regards, >> >> Jacqui >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -- >> ImageJ mailing list: http://imagej.nih.gov/ij/list.html >> > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hi Ralph,
In my experience, phase contrast images can be even more difficult to segment due to the varying gray levels and white/black granules in the cells. It's only easy if you have round cells with the halos. Phase contrast is also not really suitable for thick specimens. It's designed for thin specimens. DIC and Hoffman are both approaches for imaging thicker specimens like our 3D matrices, hence the decision to use Hoffman. Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: Ralph Common [mailto:[hidden email]] Sent: Monday, 4 May 2015 9:07 a.m. To: Jacqui Ross Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please One extra thought. A lot of the discussion in this thread involves the difficulties involved with segmenting DIC images. Is there a reason not to use phase contrast instead? The cells should be at least as visible and easier to segment. A long working distance objective such as those used with inverted scopes should let you look quite deeply into your cultures. Ralph On 5/2/2015 11:04 PM, Jacqui Ross wrote: > Hi Ralph, > > Thanks for the additional comments. We don't have a video camera available on any of our systems anymore although we do have a Dage Newvicom camera in our storeroom that could be installed I guess. I'd need to check if our current software can support it. Otherwise, I guess we could try Micro-Manager - it may have a suitable driver. > > I'm meeting with the PhD student on Monday afternoon so I will discuss the z stack approach with her again and see if it's feasible. Good to hear that Helicon Focus automatically aligns images. If it can do this with the DIC-like images, then it would be handy. > > Kind regards, > > Jacqui > ________________________________________ > From: Ralph Common [[hidden email]] > Sent: 03 May 2015 14:43 > To: Jacqui Ross > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images > - help please > > I don't want to beat a dead horse with the Z-stacks, and without > knowing any details of your research, I don't know if this will make > sense for your application. But have you considered using a video > camera to acquire the stacks? By focusing rapidly through the > specimen you could capture quite a depth in a second or two. It might > even be possible to follow the same area as it is compressed and photograph it repeatedly. > By the way, Helicon Focus does automatically align images, which is > good for me because my stage tends to drift. > > Ralph > > On 5/1/2015 2:21 AM, Jacqui Ross wrote: >> Hi Ralph, >> >> Thanks for that additional information on Helicon Focus. It is interesting that you can retouch the images prior to creating the EDF. It's an unusual approach but sounds very useful. I will take a closer look at the software as it might be useful for other applications:-). >> >> Kind regards, >> >> Jacqui >> >> Jacqueline Ross >> Biomedical Imaging Microscopist >> Biomedical Imaging Research Unit >> School of Medical Sciences >> Faculty of Medical & Health Sciences >> The University of Auckland >> Private Bag 92019 >> Auckland 1142, NEW ZEALAND >> >> Tel: 64 9 923 7438 >> Fax: 64 9 373 7484 >> >> http://www.fmhs.auckland.ac.nz/sms/biru/ >> >> >> -----Original Message----- >> From: Ralph Common [mailto:[hidden email]] >> Sent: Friday, 1 May 2015 6:18 p.m. >> To: Jacqui Ross >> Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast >> images - help please >> >> My application did not involve segmentation. I photograph a lot of spores for manual measurement or just recording what they look like, and Helicon focus does a good job of getting the spores all in focus with sharp edges. Sorry to hear your cells are moving. Brownian motion is often a problem for me. I haven't tried the programs you mention, but I have tried several others and prefer Helicon Focus. It is very fast, easy to use, and handles large images and large stacks with no problem. >> You can choose between three different methods for processing stacks depending on your type of image. In addition it has the retouching feature that lets you manually select the areas from each image you want to be in the final image, thus eliminating a lot of the artifact that results from using EDF with transmitted light microscopy. Helicon Focus is reasonably priced and has a 30 day free trial. It is not an image analysis program, but it's images can of course be processed by other programs. >> >> http://www.heliconsoft.com/heliconsoft-products/helicon-focus/ >> >> Ralph >> >> On 5/1/2015 1:54 AM, Jacqui Ross wrote: >>> Hi Ralph, >>> >>> Thanks for your suggestion. We can acquire z stacks of the cells but in this case we were compressing from the side so there was some movement in the XY direction that would confound the z stack idea because they are in a 3D matrix that also deforms. I haven't heard of Helicon Focus. We have Image-Pro Plus and NIS Elements as well and they both can create an EDF image. >>> >>> Does Helicon support segmentation as well? I can reduce the visibility of the out-of-focus cells but still can't adequately segment them out from the background. >>> >>> How did you segment your DIC images after you created the EDF image? >>> >>> Kind regards, >>> >>> Jacqui >>> >>> Jacqueline Ross >>> Biomedical Imaging Microscopist >>> Biomedical Imaging Research Unit >>> School of Medical Sciences >>> Faculty of Medical & Health Sciences The University of Auckland >>> Private Bag 92019 Auckland 1142, NEW ZEALAND >>> >>> Tel: 64 9 923 7438 >>> Fax: 64 9 373 7484 >>> >>> http://www.fmhs.auckland.ac.nz/sms/biru/ >>> >>> >>> -----Original Message----- >>> From: Ralph Common [mailto:[hidden email]] >>> Sent: Friday, 1 May 2015 5:19 p.m. >>> To: Jacqui Ross >>> Subject: Segmentation of DIC or Hoffman Modulation Contrast images - >>> help please >>> >>> Have you considered extended depth of field processing of Z-stacks? I have had good success processing DIC image stacks. There is an ImageJ plug-in for this, though I prefer to use Helicon Focus. You could, for example, take photos at 1 micron intervals, process the stack, and have many more in-focus cells in each final image before segmentation. >>> Helicon Focus has a nice "retouching" feature that would allow you to clone out any remaining out of focus cells easily before segmentation. >>> >>> Ralph Common -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Another option for stack focusing is to use the StackReg plugin to align
the images, and then the ImageJ stackfocuser plugin to generate the flattened image. We've used it on DIC images with great success. Joel Joel B. Sheffield, Ph.D Department of Biology Temple University Philadelphia, PA 19122 Voice: 215 204 8839 e-mail: [hidden email] URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* On Sun, May 3, 2015 at 7:10 PM, Jacqui Ross <[hidden email]> wrote: > Hi Ralph, > > In my experience, phase contrast images can be even more difficult to > segment due to the varying gray levels and white/black granules in the > cells. It's only easy if you have round cells with the halos. > > Phase contrast is also not really suitable for thick specimens. It's > designed for thin specimens. DIC and Hoffman are both approaches for > imaging thicker specimens like our 3D matrices, hence the decision to use > Hoffman. > > Kind regards, > > Jacqui > > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > -----Original Message----- > From: Ralph Common [mailto:[hidden email]] > Sent: Monday, 4 May 2015 9:07 a.m. > To: Jacqui Ross > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - > help please > > One extra thought. A lot of the discussion in this thread involves the > difficulties involved with segmenting DIC images. Is there a reason not to > use phase contrast instead? The cells should be at least as visible and > easier to segment. A long working distance objective such as those used > with inverted scopes should let you look quite deeply into your cultures. > > Ralph > > On 5/2/2015 11:04 PM, Jacqui Ross wrote: > > Hi Ralph, > > > > Thanks for the additional comments. We don't have a video camera > available on any of our systems anymore although we do have a Dage Newvicom > camera in our storeroom that could be installed I guess. I'd need to check > if our current software can support it. Otherwise, I guess we could try > Micro-Manager - it may have a suitable driver. > > > > I'm meeting with the PhD student on Monday afternoon so I will discuss > the z stack approach with her again and see if it's feasible. Good to hear > that Helicon Focus automatically aligns images. If it can do this with the > DIC-like images, then it would be handy. > > > > Kind regards, > > > > Jacqui > > ________________________________________ > > From: Ralph Common [[hidden email]] > > Sent: 03 May 2015 14:43 > > To: Jacqui Ross > > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images > > - help please > > > > I don't want to beat a dead horse with the Z-stacks, and without > > knowing any details of your research, I don't know if this will make > > sense for your application. But have you considered using a video > > camera to acquire the stacks? By focusing rapidly through the > > specimen you could capture quite a depth in a second or two. It might > > even be possible to follow the same area as it is compressed and > photograph it repeatedly. > > By the way, Helicon Focus does automatically align images, which is > > good for me because my stage tends to drift. > > > > Ralph > > > > On 5/1/2015 2:21 AM, Jacqui Ross wrote: > >> Hi Ralph, > >> > >> Thanks for that additional information on Helicon Focus. It is > interesting that you can retouch the images prior to creating the EDF. It's > an unusual approach but sounds very useful. I will take a closer look at > the software as it might be useful for other applications:-). > >> > >> Kind regards, > >> > >> Jacqui > >> > >> Jacqueline Ross > >> Biomedical Imaging Microscopist > >> Biomedical Imaging Research Unit > >> School of Medical Sciences > >> Faculty of Medical & Health Sciences > >> The University of Auckland > >> Private Bag 92019 > >> Auckland 1142, NEW ZEALAND > >> > >> Tel: 64 9 923 7438 > >> Fax: 64 9 373 7484 > >> > >> http://www.fmhs.auckland.ac.nz/sms/biru/ > >> > >> > >> -----Original Message----- > >> From: Ralph Common [mailto:[hidden email]] > >> Sent: Friday, 1 May 2015 6:18 p.m. > >> To: Jacqui Ross > >> Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast > >> images - help please > >> > >> My application did not involve segmentation. I photograph a lot of > spores for manual measurement or just recording what they look like, and > Helicon focus does a good job of getting the spores all in focus with sharp > edges. Sorry to hear your cells are moving. Brownian motion is often a > problem for me. I haven't tried the programs you mention, but I have tried > several others and prefer Helicon Focus. It is very fast, easy to use, and > handles large images and large stacks with no problem. > >> You can choose between three different methods for processing stacks > depending on your type of image. In addition it has the retouching feature > that lets you manually select the areas from each image you want to be in > the final image, thus eliminating a lot of the artifact that results from > using EDF with transmitted light microscopy. Helicon Focus is reasonably > priced and has a 30 day free trial. It is not an image analysis program, > but it's images can of course be processed by other programs. > >> > >> http://www.heliconsoft.com/heliconsoft-products/helicon-focus/ > >> > >> Ralph > >> > >> On 5/1/2015 1:54 AM, Jacqui Ross wrote: > >>> Hi Ralph, > >>> > >>> Thanks for your suggestion. We can acquire z stacks of the cells but > in this case we were compressing from the side so there was some movement > in the XY direction that would confound the z stack idea because they are > in a 3D matrix that also deforms. I haven't heard of Helicon Focus. We have > Image-Pro Plus and NIS Elements as well and they both can create an EDF > image. > >>> > >>> Does Helicon support segmentation as well? I can reduce the visibility > of the out-of-focus cells but still can't adequately segment them out from > the background. > >>> > >>> How did you segment your DIC images after you created the EDF image? > >>> > >>> Kind regards, > >>> > >>> Jacqui > >>> > >>> Jacqueline Ross > >>> Biomedical Imaging Microscopist > >>> Biomedical Imaging Research Unit > >>> School of Medical Sciences > >>> Faculty of Medical & Health Sciences The University of Auckland > >>> Private Bag 92019 Auckland 1142, NEW ZEALAND > >>> > >>> Tel: 64 9 923 7438 > >>> Fax: 64 9 373 7484 > >>> > >>> http://www.fmhs.auckland.ac.nz/sms/biru/ > >>> > >>> > >>> -----Original Message----- > >>> From: Ralph Common [mailto:[hidden email]] > >>> Sent: Friday, 1 May 2015 5:19 p.m. > >>> To: Jacqui Ross > >>> Subject: Segmentation of DIC or Hoffman Modulation Contrast images - > >>> help please > >>> > >>> Have you considered extended depth of field processing of Z-stacks? I > have had good success processing DIC image stacks. There is an ImageJ > plug-in for this, though I prefer to use Helicon Focus. You could, for > example, take photos at 1 micron intervals, process the stack, and have > many more in-focus cells in each final image before segmentation. > >>> Helicon Focus has a nice "retouching" feature that would allow you to > clone out any remaining out of focus cells easily before segmentation. > >>> > >>> Ralph Common > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html > -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Thanks Joel,
We don't have stacks in this case but I'll keep that in mind if the experiments are repeated. Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of JOEL B. SHEFFIELD Sent: Monday, 4 May 2015 12:45 p.m. To: [hidden email] Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Another option for stack focusing is to use the StackReg plugin to align the images, and then the ImageJ stackfocuser plugin to generate the flattened image. We've used it on DIC images with great success. Joel Joel B. Sheffield, Ph.D Department of Biology Temple University Philadelphia, PA 19122 Voice: 215 204 8839 e-mail: [hidden email] URL: *http://tinyurl.com/khbouft <http://tinyurl.com/khbouft>* On Sun, May 3, 2015 at 7:10 PM, Jacqui Ross <[hidden email]> wrote: > Hi Ralph, > > In my experience, phase contrast images can be even more difficult to > segment due to the varying gray levels and white/black granules in the > cells. It's only easy if you have round cells with the halos. > > Phase contrast is also not really suitable for thick specimens. It's > designed for thin specimens. DIC and Hoffman are both approaches for > imaging thicker specimens like our 3D matrices, hence the decision to > use Hoffman. > > Kind regards, > > Jacqui > > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > -----Original Message----- > From: Ralph Common [mailto:[hidden email]] > Sent: Monday, 4 May 2015 9:07 a.m. > To: Jacqui Ross > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images > - help please > > One extra thought. A lot of the discussion in this thread involves > the difficulties involved with segmenting DIC images. Is there a > reason not to use phase contrast instead? The cells should be at > least as visible and easier to segment. A long working distance > objective such as those used with inverted scopes should let you look quite deeply into your cultures. > > Ralph > > On 5/2/2015 11:04 PM, Jacqui Ross wrote: > > Hi Ralph, > > > > Thanks for the additional comments. We don't have a video camera > available on any of our systems anymore although we do have a Dage > Newvicom camera in our storeroom that could be installed I guess. I'd > need to check if our current software can support it. Otherwise, I > guess we could try Micro-Manager - it may have a suitable driver. > > > > I'm meeting with the PhD student on Monday afternoon so I will > > discuss > the z stack approach with her again and see if it's feasible. Good to > hear that Helicon Focus automatically aligns images. If it can do this > with the DIC-like images, then it would be handy. > > > > Kind regards, > > > > Jacqui > > ________________________________________ > > From: Ralph Common [[hidden email]] > > Sent: 03 May 2015 14:43 > > To: Jacqui Ross > > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast > > images > > - help please > > > > I don't want to beat a dead horse with the Z-stacks, and without > > knowing any details of your research, I don't know if this will make > > sense for your application. But have you considered using a video > > camera to acquire the stacks? By focusing rapidly through the > > specimen you could capture quite a depth in a second or two. It > > might even be possible to follow the same area as it is compressed > > and > photograph it repeatedly. > > By the way, Helicon Focus does automatically align images, which is > > good for me because my stage tends to drift. > > > > Ralph > > > > On 5/1/2015 2:21 AM, Jacqui Ross wrote: > >> Hi Ralph, > >> > >> Thanks for that additional information on Helicon Focus. It is > interesting that you can retouch the images prior to creating the EDF. > It's an unusual approach but sounds very useful. I will take a closer > look at the software as it might be useful for other applications:-). > >> > >> Kind regards, > >> > >> Jacqui > >> > >> Jacqueline Ross > >> Biomedical Imaging Microscopist > >> Biomedical Imaging Research Unit > >> School of Medical Sciences > >> Faculty of Medical & Health Sciences The University of Auckland > >> Private Bag 92019 Auckland 1142, NEW ZEALAND > >> > >> Tel: 64 9 923 7438 > >> Fax: 64 9 373 7484 > >> > >> http://www.fmhs.auckland.ac.nz/sms/biru/ > >> > >> > >> -----Original Message----- > >> From: Ralph Common [mailto:[hidden email]] > >> Sent: Friday, 1 May 2015 6:18 p.m. > >> To: Jacqui Ross > >> Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast > >> images - help please > >> > >> My application did not involve segmentation. I photograph a lot of > spores for manual measurement or just recording what they look like, > and Helicon focus does a good job of getting the spores all in focus > with sharp edges. Sorry to hear your cells are moving. Brownian > motion is often a problem for me. I haven't tried the programs you > mention, but I have tried several others and prefer Helicon Focus. It > is very fast, easy to use, and handles large images and large stacks with no problem. > >> You can choose between three different methods for processing > >> stacks > depending on your type of image. In addition it has the retouching > feature that lets you manually select the areas from each image you > want to be in the final image, thus eliminating a lot of the artifact > that results from using EDF with transmitted light microscopy. > Helicon Focus is reasonably priced and has a 30 day free trial. It is > not an image analysis program, but it's images can of course be processed by other programs. > >> > >> http://www.heliconsoft.com/heliconsoft-products/helicon-focus/ > >> > >> Ralph > >> > >> On 5/1/2015 1:54 AM, Jacqui Ross wrote: > >>> Hi Ralph, > >>> > >>> Thanks for your suggestion. We can acquire z stacks of the cells > >>> but > in this case we were compressing from the side so there was some > movement in the XY direction that would confound the z stack idea > because they are in a 3D matrix that also deforms. I haven't heard of > Helicon Focus. We have Image-Pro Plus and NIS Elements as well and > they both can create an EDF image. > >>> > >>> Does Helicon support segmentation as well? I can reduce the > >>> visibility > of the out-of-focus cells but still can't adequately segment them out > from the background. > >>> > >>> How did you segment your DIC images after you created the EDF image? > >>> > >>> Kind regards, > >>> > >>> Jacqui > >>> > >>> Jacqueline Ross > >>> Biomedical Imaging Microscopist > >>> Biomedical Imaging Research Unit > >>> School of Medical Sciences > >>> Faculty of Medical & Health Sciences The University of Auckland > >>> Private Bag 92019 Auckland 1142, NEW ZEALAND > >>> > >>> Tel: 64 9 923 7438 > >>> Fax: 64 9 373 7484 > >>> > >>> http://www.fmhs.auckland.ac.nz/sms/biru/ > >>> > >>> > >>> -----Original Message----- > >>> From: Ralph Common [mailto:[hidden email]] > >>> Sent: Friday, 1 May 2015 5:19 p.m. > >>> To: Jacqui Ross > >>> Subject: Segmentation of DIC or Hoffman Modulation Contrast images > >>> - help please > >>> > >>> Have you considered extended depth of field processing of > >>> Z-stacks? I > have had good success processing DIC image stacks. There is an ImageJ > plug-in for this, though I prefer to use Helicon Focus. You could, > for example, take photos at 1 micron intervals, process the stack, and > have many more in-focus cells in each final image before segmentation. > >>> Helicon Focus has a nice "retouching" feature that would allow you > >>> to > clone out any remaining out of focus cells easily before segmentation. > >>> > >>> Ralph Common > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html > -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hi Jacqui,
On 5/3/15 3:59 AM, Jacqui Ross wrote: > Hi Aryeh, > > Thanks for your suggestions. Ralph also suggested acquiring a z stack. > > The problem is that we are applying compression from the side to a 3D deformable matrix in which the cells are embedded. This means that the cells move sideways so a z stack would then require XY translation/registration to get the same cells. We could do it in that it's static compression so could hopefully work out which cells were the same ones if we had some fiducial markers (we don't!). I see that someone suggested stack registration. I just will add that you may be able to get good markers for stack registration by using the variance (or some similar edge filter) and selecting for strong features, which can produce reference marks for the alignment plugin. > If I had been involved in designing the initial experiments, I would have suggested compressing from the top, which would have resulted in less XY translation but the same deformation and we could have then used the z stack idea. I have often found that when people bring a difficult segmentation problem to me, the best way to "solve" it is to show them a better way to acquire the data (or do the experiment) so that the images are much easier to segment. Best regards, --aryeh > Kind regards, > > Jacqui > ________________________________________ > From: Aryeh Weiss [[hidden email]] on behalf of Aryeh Weiss [[hidden email]] > Sent: 02 May 2015 01:02 > To: Jacqui Ross; ImageJ Interest Group > Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please > > On 5/1/15 8:48 AM, Jacqui Ross wrote: >> Hi Aryeh, >> >> Thanks for your reply. I did try using a variance filter (the built-in one under Process - Filters - Variance) with different radii but I was unable to achieve a good result. The resultant circles were often incomplete so that when I then converted to binary, I had to do a lot of additional processing (Closing, filling holes, etc.) and then the outlines weren't very accurate. > Yes -- these methods are better at marking objects than getting accurate > boundaries. You might be able to use the inaccurate segmentation that > produces as a mask against the original variance image, which produces > reasonable arcs around your in-focus cells. >> I know that you presented on the Trainable Weka Segmentation at the ImageJ conference that I attended a couple of years ago. My notes on that weren't fantastic (I pulled them out!) but in your case you also had some fluorescence labelling to help inform the segmentation. > That problem was easier than yours (isn't it always that way?). The > texture was much better defined, and I did not have a continuum of > out-of-focus or partially out-of-focus cells. There was a DAPI channel > which I used to exclude artitfacts which were not cells (since they did > not contain nuclei). However, that will not help you, since your out of > focus cells are still cells. > > You might have an easier time here if you can acquire a z-stack (even > though it is wide-field) and use the EDF plugin or similar to have all > of your cells in-focus. That would be easier to segment. Alternatively, > if you choose to be very strict in your segmentation (ie, take only the > cells that are really sharp and textured), then I think you will be able > to variance and it relatives to get an accurate segmentation and boundary. > > Best regards > --aryeh > >> -----Original Message----- >> From: Aryeh Weiss [mailto:[hidden email]] On Behalf Of Aryeh Weiss >> Sent: Friday, 1 May 2015 5:07 p.m. >> To: Jacqui Ross >> Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please >> >> >> Try converting to 8-bit and running a variance filter >> (Process>Filters>Variance...) followed by thresholding. This will enhance the cells in focus due to their texture. >> >> --aryeh >> >> On 5/1/15 7:37 AM, Jacqui Ross wrote: >>> Hi All, >>> >>> I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. >>> >>> In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. >>> These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. >>> >>> I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. >>> >>> In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. >>> >>> The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. >>> >>> I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). >>> >>> Look forward to hearing any suggestions! >>> >>> Kind regards, >>> >>> Jacqui >>> Jacqueline Ross >>> Biomedical Imaging Microscopist >>> Biomedical Imaging Research Unit >>> School of Medical Sciences >>> Faculty of Medical & Health Sciences >>> The University of Auckland >>> Private Bag 92019 >>> Auckland 1142, NEW ZEALAND >>> >>> Tel: 64 9 923 7438 >>> Fax: 64 9 373 7484 >>> >>> http://www.fmhs.auckland.ac.nz/sms/biru/ >>> >>> >>> -- >>> ImageJ mailing list: http://imagej.nih.gov/ij/list.html >> -- >> Aryeh Weiss >> Faculty of Engineering >> Bar Ilan University >> Ramat Gan 52900 Israel >> >> Ph: 972-3-5317638 >> FAX: 972-3-7384051 >> >> >> . >> > > -- > Aryeh Weiss > Faculty of Engineering > Bar Ilan University > Ramat Gan 52900 Israel > > Ph: 972-3-5317638 > FAX: 972-3-7384051 > > -- Aryeh Weiss Faculty of Engineering Bar Ilan University Ramat Gan 52900 Israel Ph: 972-3-5317638 FAX: 972-3-7384051 -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hello Jacqui
The big problem trying to segment the DIC images are the shadow-cast effect. I work a little bit with that problem and found a interesting paper [1] which tries to eliminate this effect. Maybe after remove the shadow-cast effect the segmentation process could be more easy (in my case it was). I made an implementation of the latter method as a plugin for ImageJ and you can find here [2] if you need it. In order to use it, you need two extra images, one black (pixel value = 1) and one white (pixel value 255) (see [1] to know more of these images), with the same size of the image that need to process. Also you need to know some of the physical characteristics of the DIC filter attached to your microscope. [1] http://onlinelibrary.wiley.com/doi/10.1046/j.1365-2818.1997.2570815.x/epdf [2] http://www.ifc.unam.mx/ffm/download.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Jacqueline Ross
Hi Jacqui -
You might find these two papers, and the references therein, helpful: http://onlinelibrary.wiley.com/doi/10.1046/j.1365-2818.2000.00706.x/full http://ieeexplore.ieee.org/xpls/icp.jsp?arnumber=4761747 Best, Kurt On 4/30/2015 9:37 PM, Jacqui Ross wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. > > I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- Kurt Thorn Associate Professor Director, Nikon Imaging Center http://thornlab.ucsf.edu/ http://nic.ucsf.edu/blog/ -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
Thanks Kurt,
I came across the Arnison paper but then couldn't find an implementation of the Hilbert Transform that I could use in ImageJ. I also found this abstract http://europepmc.org/articles/pmc3635256 and tried to contact the author Paul Furcinitti to ask for more information but I didn't get a reply from him. While I can see that both of these approaches work well, I don't have any idea how to implement this into ImageJ. Neither paper mentions the software that they use so I assume they wrote their own. I might see if I can find some contact details for the authors of the more recent paper (Kuijper and Heise) as their images look very similar to ours and perhaps they might be willing to share some code. Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Kurt Thorn Sent: Tuesday, 5 May 2015 4:34 a.m. To: [hidden email] Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Hi Jacqui - You might find these two papers, and the references therein, helpful: http://onlinelibrary.wiley.com/doi/10.1046/j.1365-2818.2000.00706.x/full http://ieeexplore.ieee.org/xpls/icp.jsp?arnumber=4761747 Best, Kurt On 4/30/2015 9:37 PM, Jacqui Ross wrote: > Hi All, > > I'm helping a PhD student with analysing some Hoffman modulation contrast images of cells. She's primarily interested in changes in diameter. The cells are embedded in a 3D matrix and compression is being applied. > > In the images, there are nice cells in focus with clear boundaries, plus others which are out of focus which we don't want to measure as any measurements won't be accurate. > These kind of images are really tricky to segment as anyone who has tried, already knows. I've tried lots of different filters (edge, etc.) , FFT filtering and the Trainable Weka Segmentation but have been unable to achieve good enough results to be able to then threshold the cells automatically. > > I've come to the end of the line for now so am asking for your expert help in case anyone has some suggestions:). I note that in 2006 Monique Vasseur offered some DIC images to a PhD student called Daniel Mauch in Germany but I'm not sure if anything came of that project. There are a few papers out there (some mention Hilbert Transform, FFT) but I haven't been successful in implementing anything from those papers as yet. > > In the meantime, my solution is to use the Pseudo flat field correction plugin from Jan Brocher's BioVoxxel Toolbox (Thanks Jan!) with a very small radius (5) to flatten the background and out of focus cells while preserving the in focus cell outlines. We can then use the Cell Magic Wand (Thanks Theo!) to create selections that can be loaded into the ROI Manager and then measured. This works really well but requires that the cells be selected manually. > > The Cell Magic Wand Tool works on the colour or grayscale so we can also split the channels from the colour image if needed and use the channel image with the most contrast. > > I've attached an image in case anyone has any ideas. The image has been cropped out of a larger image so that it's not too big and it's pink because there's cell culture medium there (in case anyone was wondering..). > > Look forward to hearing any suggestions! > > Kind regards, > > Jacqui > Jacqueline Ross > Biomedical Imaging Microscopist > Biomedical Imaging Research Unit > School of Medical Sciences > Faculty of Medical & Health Sciences > The University of Auckland > Private Bag 92019 > Auckland 1142, NEW ZEALAND > > Tel: 64 9 923 7438 > Fax: 64 9 373 7484 > > http://www.fmhs.auckland.ac.nz/sms/biru/ > > > -- > ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- Kurt Thorn Associate Professor Director, Nikon Imaging Center http://thornlab.ucsf.edu/ http://nic.ucsf.edu/blog/ -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please
|
|
In reply to this post by Zian_Fanti
Hi Zian,
Thanks very much for this information on your plugin. I will definitely give it a try. I don't have the reference images you mention but maybe I can create some even without the specimen... Kind regards, Jacqui Jacqueline Ross Biomedical Imaging Microscopist Biomedical Imaging Research Unit School of Medical Sciences Faculty of Medical & Health Sciences The University of Auckland Private Bag 92019 Auckland 1142, NEW ZEALAND Tel: 64 9 923 7438 Fax: 64 9 373 7484 http://www.fmhs.auckland.ac.nz/sms/biru/ -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Zian_Fanti Sent: Monday, 4 May 2015 6:46 p.m. To: [hidden email] Subject: Re: Segmentation of DIC or Hoffman Modulation Contrast images - help please Hello Jacqui The big problem trying to segment the DIC images are the shadow-cast effect. I work a little bit with that problem and found a interesting paper [1] which tries to eliminate this effect. Maybe after remove the shadow-cast effect the segmentation process could be more easy (in my case it was). I made an implementation of the latter method as a plugin for ImageJ and you can find here [2] if you need it. In order to use it, you need two extra images, one black (pixel value = 1) and one white (pixel value 255) (see [1] to know more of these images), with the same size of the image that need to process. Also you need to know some of the physical characteristics of the DIC filter attached to your microscope. [1] http://onlinelibrary.wiley.com/doi/10.1046/j.1365-2818.1997.2570815.x/epdf [2] http://www.ifc.unam.mx/ffm/download.html -- View this message in context: http://imagej.1557.x6.nabble.com/Segmentation-of-DIC-or-Hoffman-Modulation-Contrast-images-help-please-tp5012661p5012700.html Sent from the ImageJ mailing list archive at Nabble.com. -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html -- ImageJ mailing list: http://imagej.nih.gov/ij/list.html |
«
Return to ImageJ
|
1 view|%1 views
| Free forum by Nabble | Edit this page |