Landini's Color Deconvolution module

Hi All,
I really appreciate this module
Thanks very much
I tried using the ImageJ plugin and use the results in CellProfiler- using ImageMath
This is quicker
I hope there comes an easier way to load multiple images

Could you clarify what you mean by this, since the module already allows for entering multiple stains?

Glad you like the module!
-Mark

My apologies
What I was saying about loading images is that although I can load 4images I get one set of values for all images
I would like individual images and individual data sets
I believe this has come up in the forum
I used an ImageJ plug-in that does analysis for each image in a folder and creates a summary text file for all images in the folder; each image with intensity or Haralack features.
I appreciate very much the help - especially in adding the Landini -based module
I believe it will be of great help to many who like to work with tissue microarrays
Thanks

[quote=“amaidikio”]My apologies
What I was saying about loading images is that although I can load 4images I get one set of values for all images
I would like individual images and individual data sets[/quote]

I’m still somewhat confused where you mention that you “get one set of values for all images” What value are you obtaining in this case?

The idea behind UnmixColors is that you will obtain a set of output images, one for each stain specified as input. These can then be used as input into other modules for measurement. For example, if you want the intensity or Haralick features, you would then add a MeasureImageIntensity or MeasureTexture module, or IdentifyPrimaryObjects if you want to identify specifics features within the image. Then you can use an ExportToSpreadsheet module to get the per-image measurements, which contains one row of measurement data for each image.

Regards,
-Mark

Hi ,
I should say again that CellProfiler is a great programme
I will gladly recommend to anyone; Free and a lot of helpful people
When I load 1 image ( simple Text-Exact match)
I get stain separation, then in other modules obtain image and object measurements,then ExportToSpreadsheet
When I load 4 images
I get stain separation for image#1 on screen; I do not see for image #2, #3 and #4
Output shows image and object measurements for 4 images all together; No individual image measurements

I am looking for a simple way to analyze 10 or 20 images in a folder
And to get output for image and object measurements for each image separately

I probably have to stick to doing 1 image at a time.
Thanks for all your help

If you don’t mind, would you mind uploading your pipeline so I can take a look at it? You can do so by using the “Upload attachment” tab belwo the reply text box.
-Mark

I was hoping something like List Opener (List_Opener.java) an ImageJ plugin can be available for loading images.
This loads images from a text file list of images in the same folder.
I did try another ImageJ plugin (SCSA from Univ of British Columbia Canada) for Immunohistochemical stains with similar functions; the plugin is for nuclear stains only
The plugin analyzes all files in the folder.
That kind of process will enable analysis of say 100 images per folder easily.
I do appreciate the all your help
Thanks

I think I understand now. Right now, the module allows only one input image to unmix at a time, with however many output stains you want. This is why you see only one image at time in the display window.

Where I’m getting confused is where you say “When I load 4 images.” I take this to mean either one of two things:

  • You are specifying 4 images in LoadImages. In this case, you would need to use 4 UnmixColor modules, one for each.

  • You are specifying 1 image in LoadImages for which the text pattern matches 4 images in your input folder. In this case, 1 UnmixColor module should be sufficient since CellProfiler will loop over the 4 images and record any measurements for each.

If you want to analyze 10+ images in your folder and you want to extract the same stains from each, then you would need to do option #2.

Is this what you want? I assume so since the ImageJ functionality you mention is already present in CellProfiler: you can use a LoadData module to read image lists from a .csv text file, or LoadImages to read all images in a folder that match the specified text pattern.

Regards
-Mark

Hi,
I appreciate your help
Hope not too much for you
Yes I can load images
Then use UnmixColors for each image
Then Smooth, Image Intensity , Prim ,Sec and Ter objects
Then Object measurements, Texture for EACH image
Then Export
I used six images
I hoped to get data ( intensity etc) for each image
I got ONE set of measurements for the FIRST image
I was hoping to get data for each image sequentially

I have previously attempted just
LoadImage-UnmixColors for each image -smooth- imageintensity----then Export
I get only ONE set of data NOT for each image
Is there any other way to get individual Image measurements?
Thanks in advance

Hi,
Can you upload your pipeline plus a sample image to run it on?
-Mark

Hi,
I just tried the pipeline 2010_11_24_IHC
Noticed DAB image inverted- any special reason?
I thought ApplyThreshold should be OK
Also added Secondary and Tertiary modules
To see separate nuclear and Cytoplasmic stain intensities(ImageIntensity does not provide adequate information)
No separation seen on Single image (23MB)
Do I need higher resolution to see the distinct localization?
I have images from Tissue arrays-140MB (never able to load these)- Can resize to half size

Upgraded to Snow Leopard to 10.6.5 (4G ram)
Notice more memory errors
Sometimes have re-start CellProfiller
Thanks