Demo

1. What is it?

On each slide, several different organs are present, not touching each other.
The find_objects_and_dab.groovy script was created to find and create an independent annotation for every organ on each slide. Then, it runs a pixel classifier into each newly created annotation to measure the DAB-positive area.
Since the total area of each annotation (== the total area of each organ) is known, we just have to use a spreadsheet software to process the ratio (area DAB / total area) to have the percentage of positivity within an organ.

2. How to use this script

Before starting don't forget to do the color deconvolution of your images ("Estimate stain vectors") and to apply the same one to all your batch.

We make the assumption that pixel classifiers named find-objects and find-dab are available in your project.
Open the script and adjust the settings:
- _SMALLEST_HOLE: Area of the smallest hole. Any hole smaller than this area in the detected objects will be filled.
- _SMALLEST_OBJECT: Area of the smallest object. Anything smaller will be considered as some debris and be discarded.
- _TARGET_CLASS: Class predicted by your find-objects classifier, that will also be the class of the new annotations.
Run the script for the project, it will produce an annotation this the class _TARGET_CLASS for each object.
Take a look at every slide to remove every debris that my have been caught.
You can now find a "find DAB: positive area" measurement in every annotation.

Create new classes for every type of organ you have (liver, lung, kidney, ...)
Replace the _TARGET_CLASS class by the class corresponding to the correct organ.
Now, it is possible to go in Measure > Export measurements.
Select the correct images, choose your output path, in 'export type' use 'Annotation'. Press the Export button.
In your TSV, you have a column containing the class of each ROI (liver, lung, ...), a column with the total area of the organ, and the area of positivity to the DAB staining.