SignWear AI

When Every Gesture Matters

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for SignWear AI

SignWear AI was born from a critical gap in emergency communication systems: when disaster strikes, voice-dependent infrastructure fails, leaving deaf and non-verbal users without a lifeline.

What inspired us: During a natural disaster simulation, we witnessed how emergency services remained dependent on voice-based alerts, completely excluding sign language users. We realized that accessibility isn't a feature—it's a necessity in crisis moments.

What we learned: Gesture recognition at the edge (on-device) is possible with compact neural networks. We discovered that low-latency, battery-efficient processing matters more than cloud connectivity in emergency contexts, and that co-designing with the deaf community fundamentally changes how you approach the problem.

How we built it: We created a five-layer hardware-AI system:

Sensor Integration — Flex sensors + 6-axis IMU capture hand motion with millisecond precision
Edge Signal Processing — Real-time filtering and feature extraction at the microcontroller level
On-Device AI Recognition — A CNN-LSTM model quantized to 8-bit precision, trained on ASL/BSL emergency phrases from native signers
Secure Communication — End-to-end encrypted mesh networking for reliability during infrastructure outages
Emergency Response Interface — Multi-channel alerts (text, voice, GPS) that reach emergency services, family, and support networks instantly

Challenges we overcame: