gerasa
commented
3 weeks ago I used the procedure only for one data set so I do not have a vast experience, however, I can suggest trying different parameters here:
class WatershedCellDetectionConfig { String detectionImageBrightfield = "Hematoxylin OD" double backgroundRadius = 35.0 double medianRadius = 0.0 double sigma = 3.0 double minArea = 50.0 double maxArea = 1000.0 double threshold = 0.4 double maxBackground = 2.0 boolean watershedPostProcess = true double cellExpansion = 5.0 boolean includeNuclei = true boolean smoothBoundaries = true boolean makeMeasurements = true }
to adjust to your images.
Hints how to pick parameters can be found here:
Misc. QuPath tutorial 2 – IHC analysis. https://youtube.com/playlist?list=PL4ta8RxZklWk_O_Z7K0bZlhmHtaH73vlh.
Misc. QuPath tutorial: cell detection. qupath.readthedocs.io/en/stable/docs/tutorials/cell_detection.html.
Hi there, Thank you for Celloscope, I have been trying to run the method on a 10X Visium dataset on my MacOS 14.2.1. For this, first I have installed QuPath 0.5.1 and tried to run the script cell_detection.groovy to get an estimate of number of cells at each spot. I have added the path to my coordinates file and I have added a path to an output file. However, upon adding the image and then running the script, the output shows that 0 cells are detected at each spot. I am wondering if I am doing something incorrectly, since I have not used QuPath before. Will be grateful if someone can assist me with this.
Thanks!