Adaptive Face Recognition Attendance System 🎓

An intelligent attendance system using adaptive face recognition with real-time tracking and quality assessment. Built specifically for educational institutions using SAP ID-based student identification.

🌟 Features

• Real-time Face Detection & Recognition

  • MTCNN for robust face detection
  • Pre-trained FaceNet embeddings
  • Face quality assessment
  • Multi-face tracking

• Student Management

  • SAP ID-based registration
  • Multi-angle face capture
  • Automatic database updates
  • Comprehensive student profiles

• Attendance Features

  • Real-time attendance marking
  • Subject & lecture slot tracking
  • Confidence scoring
  • Historical attendance records

🚀 Quick Start

1. Clone the Repository

   git clone https://github.com/yourusername/adaptive-face-attendance.git
   cd adaptive-face-attendance

2. Set Up Environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Configure MongoDB

   # Start MongoDB service
   mongod

4. Register Students

   python scripts/capture_student_data.py

5. Take Attendance

   python take_attendance.py

📁 Project Structure

ipd_2/
├── models/
│   ├── face_detector.py      # MTCNN face detection
│   ├── feature_extractor.py  # FaceNet embeddings
│   └── adaptive_lstm.py      # Sequence processing
├── database/
│   ├── db_manager.py        # MongoDB interface
│   └── models.py            # Database schemas
├── utils/
│   └── face_tracker.py      # Face tracking
├── scripts/
│   └── capture_student_data.py
└── data/
    └── metadata.json        # Student information

💡 Usage Guide

Student Registration

1. Update data/metadata.json with student details
2. Run capture script
3. Follow prompts to capture face data
4. Verify database entry

Taking Attendance

1. Start the system
2. Enter subject and lecture details
3. Position camera to view students
4. Monitor real-time recognition
5. Press 'q' to end session

🛠 Technical Details

• Face Detection: MTCNN (Multi-task Cascaded Convolutional Networks)
• Feature Extraction: FaceNet embeddings (Inception ResNet V1)
• Face Tracking: Custom implementation with embedding matching
• Database: MongoDB for flexible document storage
• Quality Assessment: Clarity scoring using multiple metrics

📝 Requirements

• Python 3.8+
• PyTorch 1.8+
• MongoDB 4.4+
• OpenCV
• CUDA-capable GPU (optional)

See requirements.txt for complete list.

🤝 Contributing

Contributions welcome!

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

📧 Contact

Harsh Kanani - harshkanani80@gmail.com

Smart Attendance System

A facial recognition-based attendance tracking system with an interactive Streamlit interface.

Features

• Live attendance tracking via webcam
• Process pre-recorded videos for attendance
• View and export attendance records
• Adaptive recognition using LSTM model with clarity scoring

Setup and Installation

1. Install dependencies:

   pip install -r requirements.txt
   

2. Run the Streamlit app:

   streamlit run app.py
   

Usage

Live Attendance

1. Navigate to the "Live Attendance" page
2. Enter subject name and lecture slot
3. Click "Start Camera" to begin tracking
4. Students will be recognized and marked present automatically
5. Click "Stop" when the session ends

Process Video

1. Navigate to the "Process Video" page
2. Upload a video file (.mp4, .avi, .mov)
3. Enter subject and lecture slot details
4. Click "Process Video"
5. View the results once processing completes

View Records

1. Navigate to the "View Records" page
2. Set date range, subject, and optional SAP ID filters
3. Click "Search Records" to view matching attendance data
4. Download as CSV if needed

Settings

Adjust recognition thresholds and view system information.