search

pelkmanslab / CellProfilerPelkmans

A fork of CellProfiler1.0 developed by members of Pelkmans Lab https://www.pelkmanslab.org
GNU General Public License v2.0
1 stars 1 forks source link

IdentifySecondaryIterative #11

Closed ewiger closed 9 years ago

ewiger commented 9 years ago

This error appears on analyzing 150331_HCT116_DS1_CC_F12_T0001F024L01A01Z01C01.png from /BIOL/sonas/biol_uzh_pelkmans_s5/Data/Users/Vicky/150331_HCT116_DS1_CC and is specific to the input image

Batch Error: IdentifySecondaryIterative Index exceeds matrix dimensions.
{Index exceeds matrix dimensions.

Error in IdentifySecondaryIterative (line 612)
    FinalLabelMatrixImagePre = LabelsUsed(ActualObjectsLabelMatrixImage3+1);

Error in CPCluster (line 39)
                handles = feval(ModuleName,handles);
} 
>>
tstoeger commented 9 years ago

I had already committed the patch!

Error is due to parts of standard CP code for secondary objects (which has been simply included in secondary iterative): CP assumes that there always is at least one pixel that belongs to background. If now there is no background pixel, secondary objects will be a copy of primary objects.

Ultimately I believe that it begs for a fix in the Primary modules: Previously, on other test pipelines, standard code of original CP within Discard...Objects crashed for the same reason. Not sure, why single nuclei -covering all pixels - recently start to become a problem (since having CPP).

tstoeger commented 9 years ago

If pipeline, which should be running with patched SecondaryIterative.m indeed finishes sucessfully (as it should), it might also be good to add same patch to the original IdentifySecondary module (which should have same bug)

tstoeger commented 9 years ago

Pipeline finished. Absence of background pixels is not a specific problem of IdentifySecondaryIterative. #14