Random forest classifier

---

This guide will show you how to use QuantPunc’s random forest classifier (RFC). The RFC is more generalizable than skimage’s blob counting algorithms which can lead to more accurate puncta labels.

Instructions

1. Select “Random Forest Classifier” from the Method dropdown menu.

2. The RFC requires annotations for both puncta and background. You can either:

   • Create them through Napari
   • Provide your own annotations
3. Select your annotated layer using the Select annotated layer dropdown menu.
4. Enter the integer value associated with your puncta labels in puncta_label.
5. Enter the integer value associated with your background labels in background_label.
6. Click on Train.
7. A message saying “Training complete.” should pop up indicating that your RFC has been trained.
8. Enter a value into label_color for your puncta labels and click on Label puncta to segment your puncta.