IoT · AI · App
AI Smart Stick for the Visually Impaired
Product Builder · Research Lead
Real-time object detection + OCR on Raspberry Pi, streamed to Android via WebSocket — helping visually impaired users navigate independently.
GitHub ↗APK soon
Assistive AI: Real-time CV pipeline
Edge / IoT
AI stream
Edge + companion
Hardware meets software
Device + phone bridge — sensors, on-device AI, and the companion app.
Edge device
Edge device
System story
Overview
A smart navigation assistant that detects obstacles/signs and reads text aloud in real time.
What I built
- On-device YOLO object detection for obstacles and signs
- PaddleOCR text reading with spoken feedback
- WebSocket stream to a companion Android app
- Low-latency pipeline tuned for outdoor use
- Assistive UX designed for independent mobility
What's unique
- Runs on-device (Raspberry Pi) for low-latency feedback
- Combines detection + OCR into one seamless flow
- Streams results to an Android app for a friendly experience
Why it matters
- Faster decisions → safer movement
- Works in real environments, not just demos
- Built with production thinking (real-time pipeline)
Craft
Edge stack & companion app
PythonYOLOPaddleOCRRaspberry PiAndroidWebSocket
See the full system
Explore AI Smart Stick for the Visually Impaired
GitHub ↗APK soon
GitHub ↗APK soon