RemberU is an AI-powered wearable assistant that helps users remember people they’ve met by recognizing faces and summarizing past interactions. The device combines computer vision, speech processing, and large language models to analyze and summarize conversations in real time.
Devpost: https://devpost.com/software/orientu
- Performs real-time facial detection and recognition using OpenCV, Dlib, and YOLOv8 models running on a Raspberry Pi.
- Integrates the Symphonic Labs lip-reading API to transcribe speech in noisy environments where traditional audio fails.
- Processes conversation data through the Gemini API, generating concise summaries and contextual memory notes.
- Uses Flask as a lightweight backend to handle model inference and data exchange with Firebase.
- The Flutter frontend retrieves user interaction histories from Firebase and displays personalized conversation summaries.
- Flutter
- Firebase
- Flask
- Raspberry Pi
- PyTorch
- OpenCV
- Dlib
- YOLOv8
- Symphonic Labs API
- Gemini API