Automate - creating annotation and cell detection

Hi all,

Im new at Qupath and working on ICH whole slides and applying positive cell detection.
My question is about automatization.

I have set of slides from the same sample with different markers, and my goal is to create an ROI for all images in the exactly same location of the tissue, and run the detection, so I can compare them.

The problem is, images are not exactly same. Some are little rotated, some are different in magnitude as default, etc…

What I have done so far is,
1- select baseline image and create ROI
2- copy/paste it to all images manually (shift + E) - make little manual adjustments to make it almost same location
3- run the script for project, export data, done.

What I would like QuPath to do is,
1- compare all selected images and make little adjustments automatically to bring them to same rotation/location, so I can create all annotations from script with the same location.
2- Qupath tell me where my tissue exactly is or my annotation-to-my tissue ratio, so i can calculate the location from the script regarding where to put annotation to make them all in same area.

I know very little in programming, so my queries are probably immature, but I appreciate any help.


Would need a little bit more information about just how similar your annotations are, but that sounds very similar to something already posted - but is a scripting based solution.

You also should be able to use a pixel thresholder to find the location of your tissue within an image, you might be able to use the Feret angles to estimate rotation, but it’s hard to say without seeing anything.

1 Like

Thank you so much, it worked!

1 Like