Video in the works
ClassifyAI
Marking attendance from a camera frame.
About the project
A classroom tool that takes attendance by face, tracks mood, and has an emotion-copying game.
ClassifyAI is a classroom prototype: it takes attendance by recognising faces, tracks the class's mood, and has a game where students copy emotions.
I built it in one go in April 2025 and never finished it. The emotion model is an untrained placeholder.
How it works
Add students
Upload the class list from Excel.
Take attendance
A camera frame is matched against students' faces.
Confirm
Voice confirmation of who was marked.
See it
Attendance and analytics charts on the dashboard.
Under the hood
- Backend
- FastAPI with face_recognition, OpenCV and TensorFlow.
- Frontend
- React with MUI and Chart.js.
- Setup
- Postgres and Docker Compose.
What I did
- A FastAPI backend for students, attendance, analytics and the game
- Attendance by face recognition, and voice confirmation
- Student upload from Excel
- A React dashboard with charts
- Docker setup and a database schema
Screenshots
Screenshots in the works
Real screens from ClassifyAI go here.
To be clear
The emotion model is an untrained placeholder, so mood tracking can't work as built, and face data is kept in memory, not saved.