I want to build an AI that plays a simple android game.
The game is just a one at a time object falling, some times at an angle. The AI needs to recognize the object and to decide whether to swipe left, swipe down, or click on it. The background is changing some times, but the object falling is always on top.
There are 44 different assets and I have the original full resolution PNG of the objects.
How should I approach this?