Done by Zhonglin Niu (zn23), and Yuese Li (yl77).

"ClassCheckIn uses Raspberry Pi and a camera to instantly recognize students by their face as they enter class, automating attendance with just one photo per student."

1. Copy the .env.local.sample file to .env.local and fill in the required variables.

2. Copy the face_recognition/.env.sample file to face_recognition/.env and fill in the required variables.

3. On local machine, run the following commands in the root directory (I used python3.12):

   cd class-checkin
   python3 -m venv venv
   source venv/bin/activate
   pip install -r face-recognition/requirements.txt
   npm install
   npm run dev

4. On the Raspberry Pi, run the following commands:

   cd RPI
   python3 cam_capture.py
   python3 mqtt_client.py

We were frustrated with time-consuming attendance methods like scanning QR codes, logging into Moodle, or professors calling roll. We wanted to create a seamless attendance system that wouldn't disrupt valuable class time.

ClassCheckIn uses facial recognition to automatically mark students present as they enter the classroom.

We developed the web interface using React+Vite, captured video streams from a Raspberry Pi camera, and implemented Python-based facial recognition that sends data to a cloud database for the web interface.

• Due to the Raspberry Pi's limited memory (1GB), facial recognition had to be performed locally
• Optimizing facial recognition speed while maintaining accuracy
• Implementing reliable MQTT communication between components for real-time feedback
• Integrating multiple systems (web interface, recognition engine, and hardware feedback) into a cohesive solution

Our system confirms successful recognition through visual and audio feedback - green LED flashes and short buzzer sounds via MQTT protocol. The entire recognition process takes less than one second, and all attendance data is securely stored in our cloud database.

Working on ClassCheckIn taught us how to integrate hardware and software components seamlessly. We gained valuable experience in facial recognition algorithms, real-time data processing, and creating responsive user interfaces. The project also strengthened our skills in database management and IoT communication protocols, particularly MQTT for device coordination.

We plan to enhance our database architecture to enable professors to track student attendance throughout an entire semester and export the data for their records.