SMS | Devpost

Home Page
Student Interface
Student Attendance
Online Quizzes & Homework
Parents Interface
Grades
Creating Quizzes

Inspiration

This project began as our graduation thesis—a technical exploration of AI-driven attendance systems. But as we saw this opportunity (CodeForSudan) and the war ravaged Sudan, displacing millions and crippling education, we asked ourselves: How can we adapt this to help Sudanese students survive this crisis?

In Sudan, the devastating impact of war has disrupted education, leaving schools struggling with outdated systems, poor record-keeping, and frequent student displacement. Many institutions still rely on manual attendance tracking, which is inefficient and prone to errors—especially in crisis situations where students may relocate suddenly or face irregular attendance due to safety concerns.

What it does

The School Management System (SMS) is a web-based platform that:

Uses facial recognition for student login and automated attendance
Allows instructors to upload materials, assign quizzes, and track grades
Enables students to view their progress and submit assignments
Offers a parent portal to monitor their child's academic journey
Provides admins full control over users, courses, and system data

How we built it

We used a full-stack approach integrating both software engineering and machine learning:

Frontend: HTML, CSS, JavaScript
Backend: Python with Flask/Django
Database: MySQL / PostgreSQL
Face Recognition: TensorFlow/Keras with OpenCV
Version Control & Deployment: Git & GitHub Our custom CNN-based facial recognition model was trained on a dataset of over 17,000 images, achieving over 95% validation accuracy.

Challenges we ran into

Building an accurate face recognition model with limited subject diversity
Designing responsive dashboards tailored to different user roles
Ensuring secure and scalable data handling for sensitive educational information
Real-time synchronization between facial recognition and attendance logging

Accomplishments that we're proud of

Achieved over 95% validation accuracy with our custom CNN
Developed a fully functional, multi-role school management system
Integrated machine learning seamlessly into a real-world web application
Successfully simulated real-time attendance using live webcam input

What we learned

Applying deep learning models (CNNs) in practical, user-facing applications
Building secure, scalable backend systems for real-time data
Managing large datasets for model training and validation
Importance of clean UI/UX when dealing with multi-role platforms

What's next for School Management System

Enhance facial recognition to work across age and time (face aging models)
Integrate advanced analytics for student performance insights
Develop a mobile version of the platform for broader accessibility
Expand role support (e.g., librarian, counselor) and integrate scheduling features