Intensity quantfiication

Hi everybody,

I have a huge problem with intensity quantification since switched to CellProfiler 2.0. It seems that in low intensity pictures (B02-B06) when you use the 2.0 version the picture intensity is pushed to the limit not only in the preview but also in the one that is used for calculation. This results in low intensity pictures giving higher readings than pictures with a stronger intensity (B07-B11). I have attached the pipeline,as well as the result files for a run with 1.x version (where everything seems to be quantified correctly) and with the 2.0 version. Unfortunately I am not able to attach the sample pictures, I can provide them if necessary. I would really appreciate if someone could help me solve this problem

Thanks
DefaultOUT_old.mat (211 KB)
DefaultOUT_new.mat (21.8 KB)
version-testPIPE.mat (650 Bytes)

Yes, if possible, please provide an example of the raw input images. if you can’t upload it to the forum, we recommend using YouSendIt, or Picasa if you have an account there.
-Mark

Thanks a lot for the reply, the images can be downloaded via the following link

The files are (follow link to download):

siRNA_screen_3A_d7_UCP1_B02_04_w530.tif cifex.ethz.ch:443/?fileId=6128& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B11_12_w530.tif cifex.ethz.ch:443/?fileId=6119& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B10_12_w530.tif cifex.ethz.ch:443/?fileId=6120& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B09_10_w530.tif cifex.ethz.ch:443/?fileId=6121& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B08_02_w530.tif cifex.ethz.ch:443/?fileId=6122& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B07_10_w530.tif cifex.ethz.ch:443/?fileId=6123& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B06_04_w530.tif cifex.ethz.ch:443/?fileId=6124& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B05_03_w530.tif cifex.ethz.ch:443/?fileId=6125& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B03_03_w530.tif cifex.ethz.ch:443/?fileId=6126& … wolfrum.ch
siRNA_screen_3A_d7_UCP1_B04_13_w530.tif cifex.ethz.ch:443/?fileId=6127& … wolfrum.ch

Expiration: The files will be removed on 2010-10-03 at 23:59:59

For downloading you have to enter the following password: t/7i[ocF?Rh*XN@eIv%

Link to login to the system: cifex.ethz.ch:443/?user=christian%40wolfrum.ch

Hi,
The problem seems to be with your vendor’s TIF format. The TIF format embeds several pieces of information in your image file including one, MaxSampleValue, that should indicate the value that is output by the vendor’s A/D acquisition device at saturation (typically 4095 for a 12-bit converter or 65535 for a 16-bit one). It appears that your vendor has a different interpretation for this field.

I’m going to add a checkbox to LoadImages, LoadSingleImage and LoadData to handle this case - the default will be to scale images, but if you uncheck the box, it will not use the MaxSampleValue to scale. Furthermore, the MaxSampleValue is useful to have as a measurement - we have two separate requests to retain this value, so I’ll add that as a measurement that might be used in a downstream module (for instance ImageMath).

I’ll post here once I have the changes in and tell you how you might pick up a version that supports them. Which platform do you use?

–Lee

The “Rescale intensities?” checkbox has been implemented for awhile now in our code. Until we release a new version on our website, you can try our newest “bleeding edge” version here, which will have the new checkbox:
cellprofiler.org/cgi-bin/trunk_build.cgi

-David