What if your FlipperZero could scan the scene visually to match payloads to target devices?

In this example project, we trained a tiny image classifier to differentiate between iPhones and RFID card readers from an ESP32-CAM feed. We show how to create custom model optimized to run in real-time on your MCU.

We're planning to integrate this example with the Mayhem project: https://github.com/eried/flipperzero-mayhem

What kind of devices would you want to identify? Feedback/thoughts are welcomed!