Taking attendance is one of those everyday tasks that feels small but consumes a surprising amount of time—whether in classrooms, training sessions, or workplaces. For years, I watched teachers and staff spend 10–15 minutes marking attendance manually, and proxy attendance was still a common problem.
Why This Project?
The manual attendance process is slow, prone to errors, and easy to misuse.
I wanted to build a system that:
- Marks attendance instantly
- Eliminates proxy attendance
- Is easy for students and staff to use
- Runs directly in the browser without installing software on every device
Face recognition felt like the perfect solution—fast, accurate, and user-friendly.
How the System Works
The flow is designed to be simple for both students and administrators.
For Students
- Register using a webcam (one-time setup)
- Show your face to mark attendance
- Attendance is recorded within seconds
- Only one attendance per day is allowed
For Admin / Staff
- Register and manage students/employees
- View daily and monthly attendance
- Generate downloadable Excel reports
- Access a dashboard with analytics and insights
The entire system works through a web browser, making deployment and access very straightforward.
Technology Used
I kept the tech stack lightweight, scalable, and easy to maintain:
- Python (Flask Framework) – for backend and routing
- OpenCV (Haar Cascade Classifier) – for face detection
- MySQL – for securely storing user data and attendance
- HTML/CSS – for a clean, minimal frontend
- Pandas + OpenPyXL – for generating Excel reports
Face Recognition Flow
- A user’s face is captured during registration.
- It is converted into numerical encodings.
- These encodings are saved in the database.
- During attendance, the system captures the live face.
- It compares both encodings.
- If they match → attendance is marked automatically.
Key Features
- Face-based login and attendance
- Role-based access (Admin & Student)
- Daily, monthly, and custom-date reporting
- Excel report generation
- Error handling for multiple faces
- Fast processing and clean UI
Performance & Results
After multiple tests, the results were strong:
- Attendance marking speed: 5 seconds (average)
- Recognition accuracy: ~95%
- Zero chance of proxy attendance
- Reports generated in 2–3 seconds
- Users were able to adapt quickly with no training required
Future Enhancements
I plan to extend the system further with:
- Mobile app support
- Email/SMS attendance notifications
- Advanced analytics with charts
- Mask detection
- Cloud-based hosting and database integration
How to Try It
- Install Python (3.10+ recommended)
- Install XAMPP
- Install dependencies using pip
- Run the Flask app
- Access the system at: https://updategadh.com
Default Admin Login
- Username: admin
- Password: admin123
(recommended to change before deployment)
Download :-Click here
