It’s amazing how little hardware is necessary to get started with programming custom image recognition systems. Today we have a fantastic project to share by engineer Mohammad Reza Sharifi who’s using a Raspberry Pi with OpenCV to read hand gestures and trigger custom responses.
Image recognition isn’t something new on the Raspberry Pi; we’ve even dabbled in recognizing faces ourselves. However, this is one of the few examples we’ve seen in which hands are mapped out and read for specific gestures.
The system operates in tandem with Tensorflow to train the gesture models. Finally, Python brings everything together by reading the gestures as input and triggering a response. To get a closer look at the source code used in this project, check out the official page at GitHub.
In a demo video shared by Sharifi, the system drives a robotic car. He’s seen using various signs to trigger specific commands like moving forward, stopping, or determining which direction for the vehicle to steer.
Check out the project in action on Sharifi’s official YouTube channel, and be sure to follow him for more cool creations.