Muggle Wand Training

Muggles usually can't cast spells. Now, they can through the power of edge AI computer vision inference. Interact with both the physical and digital world through wand gestures and voice commands.

Comment

Inspiration

We are fans of the Harry Potter books and movies, and wanted to bring the magic to life using Qualcomm's hardware.

What it does

Detects wand gestures and voice commands to detect certain spells being cast. Our web app and Arduino Uno Q microcontroller react based on the spell cast in the digital and physical world, respectively.

How we built it

As a team, we brainstormed different ideas both technically and creatively. We consulted mentors and got user feedback from fellow hackers. We also used agentic LLMs to rapidly prototype.

Challenges we ran into

Detecting spells via just computer vision didn't work too well, so we also used voice recognition in addition to make spell detection more accurate. Sending messages from the app to both the Flask web application and the Arduino Uno Q via MQTT also proved to be difficult.

Accomplishments that we're proud of

The fact that our project spans an Android app, a web application, and an embedded system is something we are very proud of. We are also proud of how creative our idea is.

What we learned

We learned how to use an Arduino Uno Q, including its strengths and weaknesses. Also, Kotlin app development on Android.

What's next for Muggle Wand Training

More spells. More interactivity. More accurate spell detection.

Share this project:

Updates