I am trying to classify micropillars using CPA.
Initially, I segment the images using CellProfiler, and identify the micropillars as primary objects. I have made sure that, when creating the CPA Properties file, I select these objects (micropillars) to be used for locations. I belive CellProfiler is correctly identifying all micropillars as my primary objects, because I have an output image where these are marked. I have also looked at the CellProfiler output csv file where these primary objects are identified, and each image seems to have the expected number of micropillars. I process around 25 images on CellProfiler, and each image has 300-400 micropillars.
When loading the Exp file to CPA and using the Classifier module, I am able to create a training set with some objects from all images.
However, when opening some individual images from the set, I am not able to “click” on all micropillars, i.e. CPA only considers the first 20 objects, so I can see micropillars from the source image, but they are not marked as the primary objects from CellProfiler. For some reason, I notice it happens with the last images to be processed. So, if I am working with 25 images, the first 5 are ok (the CPA final count is around 300 objects), but for the rest, CPA only counts 20 objects.
If anyone could help me, I would be very thankful!
Have a nice day!