Segmentation of yeast cells. Please help!

[Giuseppe Lucarelli]

Hi there,

I am currently stuck on cell segmentation of my images. I deal with yeast cells and I am trying to get cell ROI using two consecutive filters on confocal fluorescence images of yeast cells. The cell express a fluorescent tag which localises close to the cell wall but it is not a continuous and homogeneous signal around the cell, it can be dotty creating voids in the cell "perimeter". I have attached an example image (1024x1024, 16-bit). WT_24h_-N_green.tif

So far I apply the filters (in this order):
1) Unsharp Mask,
2) Variance

The parameters for these filters need to change in order to get most cells out of the different images.

The problem is that I cannot come up with a strategy to find out which parameters to feed into the filters depending on some property of the image to be processed.
I could loop the macro over a range of parameters and then use the parameters that lead to the highest number of cells being identified but this will take a long time especially because I have > 2000 images to process.

Do you think I am using the right approach/filters? Do you have any suggestions?
Should I try to segment cells from the brightfield image?

Thanks a lot for your help.

[Thomas Boudier]

Hi Guiseppe,

Usually for yeast cells, I use band pass filtering since almost all
cells have same area (FFT/BandPass). It will help enhance the cells and
get rid of the variation of intensities. But then what is your question
? Number of cells ? Sizes ?
Best,

Thomas


On 09/12/14 11:52, giuseppe3 wrote:

> Hi there,
>
> I am currently stuck on cell segmentation of my images. I deal with yeast
> cells and I am trying to get cell ROI using two consecutive filters on
> confocal fluorescence images of yeast cells. The cell express a fluorescent
> tag which localises close to the cell wall but it is not a continuous and
> homogeneous signal around the cell, it can be dotty creating voids in the
> cell "perimeter". I have attached an example image (1024x1024, 16-bit).
> WT_24h_-N_green.tif
> <http://imagej.1557.x6.nabble.com/file/n5010855/WT_24h_-N_green.tif>
>
> So far I apply the filters (in this order):
> 1) Unsharp Mask,
> 2) Variance
>
> The parameters for these filters need to change in order to get most cells
> out of the different images.
>
> The problem is that I cannot come up with a strategy to find out which
> parameters to feed into the filters depending on some property of the image
> to be processed.
> I could loop the macro over a range of parameters and then use the
> parameters that lead to the highest number of cells being identified but
> this will take a long time especially because I have > 2000 images to
> process.
>
> Do you think I am using the right approach/filters? Do you have any
> suggestions?
> Should I try to segment cells from the brightfield image?
>
> Thanks a lot for your help.
>
>
>
> --
> View this message in context: http://imagej.1557.x6.nabble.com/Segmentation-of-yeast-cells-Please-help-tp5010855.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
   /***************************************************************/
      Thomas Boudier, Associate Professor, UPMC,
      Université Pierre et Marie Curie, Paris, France.
      BioInformatics Institute (BII)/IPAL, Singapore.
/**************************************************************/

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Giuseppe Lucarelli]

Hi Thomas,

thanks for your message.

I am trying to get the max number of cells out of each image to then add it to the ROI manager for further processing.

I have never used a band pass filter and have no idea on how it works. I'll read some more about today.
Thanks again.

Giuseppe