Signyfy
An AI-powered smart glasses prototype that translates sign language into speech in real time by combining computer vision, machine learning, and wearable hardware.
PythonOpenCVMediaPipeTensorFlowRaspberry PiAzure TTS
Role
Computer vision pipeline and wearable hardware integration (team project)
Timeline
MorganHacks 2024
Problem
Communication between sign language users and non-signers often depends on another person or device already understanding sign language, limiting independent real-time conversation.
Approach
Built a wearable prototype that captured hand gestures through an embedded camera, extracted pose features with MediaPipe, recognized gestures with a TensorFlow model, and converted the resulting text into speech on a Raspberry Pi.
Results
- Won first place and the $4,000 top prize at MorganHacks 2024.
- Integrated camera input, gesture recognition, machine learning, and text-to-speech into one hardware prototype.
- Established a working data pipeline and responsive hand-tracking foundation during the hackathon.
Impact
- Recognized as the top overall project among the hackathon submissions.
- Demonstrates end-to-end computer vision work beyond a standalone model or interface mockup.
- Combined software and physical hardware under a two-day delivery constraint.
Gallery