Apppublic

youssefnaiem/hjbfhjeb_cjkdwnbj

sourceHugging Faceupdated 7mo agoView on Hugging Face

0likes

App README

🎓 Smart Face Emotion-Based Attendance System

A real-time AI-powered attendance system that uses Face Recognition + Facial Emotion Detection to automatically mark attendance through a university-style web portal.

This project integrates Deep Learning, Computer Vision, and Web APIs to deliver a fully automated, contactless attendance solution.

🚀 Features

✔ Real-time face detection using MediaPipe ✔ Face recognition using FaceNet embeddings + SVM classifier ✔ Facial emotion recognition using CNN trained on FER-2013 dataset ✔ Automatic attendance marking with Date, Time, Emotion & Status ✔ Clean University Portal UI (HTML, CSS, JavaScript) ✔ Backend powered by FastAPI ✔ Camera opens only on user action ✔ Attendance stored in CSV (can be extended to DB)

🧠 System Architecture User → Web Portal → FastAPI Backend ↓ Webcam Capture ↓ Face Detection ↓ Face Recognition (FaceNet) ↓ Emotion Detection (CNN) ↓ Attendance Marked (CSV)

🛠 Tech Stack 🔹 Backend

Python 3.11

FastAPI

OpenCV

MediaPipe

TensorFlow / Keras

FaceNet (keras-facenet)

Scikit-learn

Pandas

🔹 Frontend

HTML5

CSS3

JavaScript (Fetch API)

🔹 Deep Learning

FaceNet (128-D embeddings)

CNN (FER-2013 emotion dataset)

📸 How Attendance Works

1️⃣ User opens the web portal 2️⃣ Clicks “Mark Attendance” 3️⃣ Camera opens 4️⃣ System detects:

Face

Identity

Emotion 5️⃣ Attendance is marked as Present 6️⃣ Camera auto-closes 7️⃣ Popup confirms attendance

📊 Attendance Format (CSV) Name Date Time Emotion Status Lakshya 2025-12-15 15:30:12 Happy Present Karan 2025-12-15 Absent

🎯 Emotion Classes Supported

Happy

Sad

Angry

Neutral

Surprise

Fear

Disgust

🧪 Model Training Emotion Model

Dataset: FER-2013

Model: CNN

Framework: TensorFlow / Keras

Face Recognition

Embeddings: FaceNet

Classifier: SVM

🌟 Future Enhancements

🔐 Database integration (MySQL / Firebase)

📊 Attendance analytics dashboard

🎥 Live emotion overlay

📱 Mobile-friendly UI

🏫 Multi-class support

🔒 Admin authentication

🧑‍💻 Author

Lakshya Srivastav Computer Vision & Deep Learning Project

⭐ Acknowledgements

FER-2013 Dataset

MediaPipe by Google

FaceNet Research Paper

TensorFlow & OpenCV Communities

📌 License

This project is licensed for academic and educational use.