I’ll notify @Sofia here but I’m not sure she’s still active on the forums. She’d be the best person to ask about this.
I can tell you from experience training an image recognition AI that it’s complicated work. You can read about my attempts here: Image Recognition & Classification using Artificial Intelligence. It took me hours of trial and error and reading documentation to get it right. And even after all that, I couldn’t afford to pay for what I really needed to make an app.
The built-in image recognition in Thunkable is pretty basic but you could start with that and see if it’s enough for you: Camera Blocks - Thunkable Docs.