-
AI-Powered Smart Attendance & Emotion Analysis Dashboard – Real-time Face Recognition, Emotion Tracking, and Attendance Analytics System
Inspiration
In our current education ecosystem, roll calls, ID cards, and signatures still determine attendance, which wastes teaching time and allows proxy/buddy attendance. Survey of students suggest that 30–40% of college students in India have some experience with proxy attendance, which gives rise to discipline and inaccurate academic reports.
With Digital India and AI adoption at the highest levels, I wanted to solve a simple, widespread problem: ✅ Save Classroom Time ✅ Increase Fairness (and Decrease Biasness) ✅ Improve Student Accountability ✅ Modernize Campus Systems with AI
After seeing teachers spend 10–15 minutes every day on attendance logging, I thought some technology could free that time for real learning.
What it does
EduMark AI is a live face recognition attendance application for schools, colleges, and offices.
It:
Instantly recognizes faces
Automatically marks attendance
Prevents proxy / false entries
Creates time-stamped logs
Displays live recognition results on screen
Securely stores attendance
Works with VGA and CCTV streams
In summary:
Your face = your ID Attendance, in seconds, with full authenticity.
How we built it
Python backend
OpenCV for live video processing
Face Recognition (dlib model) for embedding and matching
NumPy & Pandas for data handling
Flask for web UI display
CSV/SQLite for storage Live Camera → Face Detection → Face Encoding → AI Matching → Attendance Log → Web Dashboard Optimized for real-time accuracy and low-latency processing.
Challenges we ran into
Variations in face quality (such as lighting, angle, and crowding)
How to maintain real-time performance on any normal laptop hardware
Create a simple UI that teachers can use without training
Develop a clean dataset of facial encodings
Each challenge drove us to test, iterate, and improve recognition confidence and system stability.
Accomplishments that we're proud of
Complete functioning AI attendance system from recorded class lessons
Over 95% recognition accuracy in experiments
Automatic detection and logging
Included web dashboard
Developed working prototype (not a research concept)
We built it, made a demo video and wrote documentation
This project demonstrates we can produce practical deployable AI in education.
What we learned
A comprehensive understanding of face embeddings and feature vectors
The integration of real-time machine learning and computer vision
Tuning performance and constructing datasets
Product-minded – designing for real people and not just theory
Storytelling and pitching a scalable AI solution
What's next for EduMark AI – Smart Identity & Attendance
Detecting liveliness (anti-spoofing) to prevent phone/video cheating
Cloud-based admin dashboard for management
Mobile attendance application for teachers
Integration of multiple cameras in campus installs
Automated report generation & dashboards for insights.
Pilot deployment in schools and colleges for evaluation.
Vision: To make EduMark AI the standard attendance system in campuses in India for the future of AI enabled smart institutions.
Log in or sign up for Devpost to join the conversation.