🧠 CogniCare - Companion App for Memory Support

🏆 Social Impact Track Winner @ MHacks 16

Overview

🚀 This project, built at the University of Michigan, is designed as a companion for individuals facing memory recollection challenges, including those at risk of Alzheimer's disease or Dementia.

🤖 Leveraging AI-powered memory analysis and cognitively stimulating brain exercises, we're supporting memory health and allowing patients to remember their lives by reliving their memories.

Key Features

• 🎮 Interactive Brain Exercises: Generative AI-powered games that stimulate cognitive functions, intelligently generated from your daily experiences.
• 🤖 AI-Powered Memory Access: Your memories are analyzed, stored, and accessible through natural language queries.
• 👩‍💻 Powerful, Personalized Experience Recollection: The mental activities are tailored for users with memory recollection issues; they can watch their own memories to remember important events.
• 📊 Data-Driven Insights: Analytics to track health status and provide valuable feedback for users and caregivers.

Running Locally

Cloning the repository the local machine.

git clone https://github.com/ishaan1013/mhacks

Add your OpenAI API key (with GPT-4 access)

• Required in backend/app.py and backend/gpt_4.ipynb

Setup the backend API

• 🪐 Preprocess video data with the Jupyter notebook files, which perform the following:

  • Converting videos into short fragments
  • Transcribing audio & analyzing frames chosen from video fragments using ORB analysis
  • Performing facial recognition
  • Generating questions, names, etc. for brain exercises

• 🐍 Run the flask app, which will be available at http://127.0.0.1:5000.

flask run

Running the frontend application

npm install

Then, run the application in the command line and it will be available at http://localhost:3000.

npm run dev

Acknowledgments

• 🙏 Memory analysis inspiration (with permission) from Rajan's team's Hack the North project in 2023.