ABSTRACT
This Python Flask web application automates student attendance using face recognition, integrating OpenCV and the face_recognition library for biometric identification. Designed for educational institutions, it features role-based access with Flask-Login, distinguishing admins (who manage attendance) from regular users. During registration, students capture images via webcam, which are processed using Haar cascades to detect and store facial encodings. The system trains a model by converting these images into numerical encodings, saved as NPZ files for efficient matching. Attendance is recorded by comparing live camera feeds or uploaded images against the trained dataset, leveraging Euclidean distance to identify students. Results are logged in timestamped CSV files organized by subject (e.g., Math, Science) and date, with options to view or download reports. The backend uses MySQL for user authentication, with password hashing for security, while the frontend employs JavaScript for real-time face detection during registration and attendance. Admins can trigger training, mark attendance by subject, and generate reports, while the system handles edge cases like unrecognized faces. Configuration files manage paths for uploads, training data, and attendance records, ensuring modularity. Dynamic subject lists are derived from predefined options and attendance files, with dates auto-populated for filtering. Secure practices include sanitized file uploads, login-protected routes, and admin-only access for critical operations. The app follows MVC architecture, with Flask routes handling requests, Jinja templates rendering views, and Python logic processing face data. APIs return JSON responses for seamless frontend interactions, such as live face detection during camera capture. Error handling addresses missing training data or failed face detection, providing user feedback. Scalability is maintained through filesystem-based student records, though future versions could transition to a full database. The system demonstrates practical AI integration in web apps, balancing performance (with NPZ-based encodings) and usability (with real-time feedback). Its open-source stack—Flask, OpenCV, and lightweight libraries—makes it adaptable for schools with limited infrastructure.


OBJECTIVES
1. Automate Attendance Tracking
To develop a contactless system that accurately records student attendance using facial recognition technology, eliminating manual processes and reducing administrative workload.
2. Ensure Accurate Identification
To implement robust face detection of Haar cascades and recognition algorithms capable of handling variations in lighting, angles, and appearances while minimizing false positives/negatives.
3. Enhance Security & Prevent Fraud
To create a tamper-proof attendance system that prevents proxy attendance through biometric authentication and maintains secure records with role-based access control.
4. Provide Accessible Data Management
To design an intuitive interface for viewing, organizing, and exporting attendance data by subject/date, with downloadable reports for administrative use.
5. Optimize Institutional Efficiency
To reduce time spent on attendance-related tasks by 70% while improving data accuracy, enabling educators to focus on teaching rather than administrative duties.