Smart Attendance System

Overview

A web-based smart attendance system using face recognition. It leverages a Flask backend, MediaPipe for face detection and alignment, and ONNX models for efficient face embedding generation.

Key Features

• Web Interface: Easy-to-use browser-based interface for registration, recognition, and reporting.
• Face Registration: Register students by capturing their face via webcam directly from the browser.
• Face Recognition: Real-time face recognition from the webcam feed to mark attendance automatically.
• Attendance Management & Reports: View attendance records with advanced filtering and export them to CSV, JSON, or Excel.
• Database Storage: Uses SQLite and SQLAlchemy to robustly store user profiles, face encodings, and attendance logs.

Technologies Used

• Backend: Python 3.11+, Flask, Flask-SQLAlchemy, Pydantic
• Computer Vision: mediapipe (face detection/landmarks), opencv-python (image processing), numpy, onnxruntime (Mobile-ArcFace/MobileFaceNet for embeddings)
• Database: SQLite
• Package Management: uv

How It Works

1. Registration: Users navigate to the Registration page and capture their face. The system generates a 512-dimensional embedding and stores it in the database.
2. Recognition: The Recognition page captures the live feed via webcam. It extracts faces and compares the embeddings against the registered dataset.
3. Reporting: Successfully identified faces are securely logged into the SQLite database and can be exported as reports.