This short tutorial describes, how to to train Power Apps AI model to visually recognize objects on pictures.

In this case, model is going to distinguish parrots, guinea pigs and hamsters from each other.

How to start?

In Power Apps select new model and Object Detection.

Type some name and continue.

If you don´t need anything special, keep the selection for Common objects.

Write names of groups / categories, you are assigning into. Names are not important.

Now you need at least 15 images for each category. 15 guinea pigs, 15 hamsters, 15 parrots. 

Now check all images and tag objects. You can tag multiple objects on one image. Rectangles showing objects are created by Power Apps, but you can modify them manually. 

You can not continue, if you don´t have at least 15 images for each category.

Confirm and “Train”.

Now it needs few seconds to learn what is what.

Click on it and try how it works. Simply try “Quick test” and load some image. Model is supposed to recognize, what is there.