# Interactive object classification in Napari

In this exercise we will train a [Random Forest Classifiers](https://en.wikipedia.org/wiki/Random_forest) for classifying segmented objects. 
We will use the napari plugin [napari-accelerated-pixel-and-object-classification](https://www.napari-hub.org/plugins/napari-accelerated-pixel-and-object-classification).

## Getting started

Open a terminal window and activate your conda environment:

```
conda activate devbio-napari-env
```

Afterwards, start up Napari:

```
napari
```

Load the "Blobs" example dataset from the menu `File > Open Sample > clEsperanto > Blobs (from ImageJ)`

We furthermore need a label image. You can create it using the [pixel classifier trained earlier](machine_learning:pixel_classification) 
or using the menu `Tools > Segmentation / labeling > Gauss-Otsu Labeling (clesperanto)`.

## Object classification

Our starting point is a loaded image and a label image with segmented objects. The following procedure is also shown in [this video](apoc_object_classification.mp4).

![](apoc21.png)

Add another label image. Rename the label image e.g. to `Label class annotation` to not mix it up with the other.
![](apoc22.png)

Activate the `Brush tool`.
![](apoc23.png)

Put small dots with label `1` into small roundish objects (for training purposes: really just the smaller ones).
![](apoc24.png)

Increase the `label` to `2`.
![](apoc25.png)

Draw a line throught the larger elongated objects in the center of the image.
![](apoc26.png)

Start the object classification tool from the menu `Tools > Segmentation post-processing > Object classification (APOC)`
![](apoc27.png)

In this user-interface, activate the `shape` checkbox.
![](apoc28.png)

Select `image`, `labels` and `annotation` like this:
![](apoc29.png)

Click on `Run`. After a second a new labels layer with brown / blue annotated objects should appear. Some larger round objects will be blue unintentionally.
![](apoc30.png)

Hide the newly created classification layer.
![](apoc31.png)

Select your annotation layer.
![](apoc32.png)

Annotate some more roundish objects, this time the larger ones.
![](apoc33.png)

Train the classifier again.
![](apoc34.png)

If you are happy with the trained classifier, copy the file to a safe place. When training the next classifier this one might be overwritten.

## Extra exercise
Retrain the classifier so that it can differentiate three different classes:
* Small round objects
* Large round objects
* Large elongated objects