Face Recognition Attendance System

This is a simple web-based face recognition attendance system implemented using Flask and OpenCV.

Overview

The system captures real-time video feed from the webcam, detects faces using Haar cascade classifier, and recognizes them based on a pre-trained KNN classifier. Users can add new faces to the system, and the attendance of recognized individuals is logged in a SQL Database.

Prerequisites

• Python 3.x
• OpenCV
• Flask
• NumPy
• scikit-learn

Installation

1. Clone the repository:

   git clone https://github.com/scorpionTaj/face-recognition.git

2. Install dependencies:

   pip install -r requirements.txt

Usage

1. Run the Flask application:

   python app.py

2. Open a web browser and go to http://localhost:5000/.

3. Click on the "Prendre la Présence" button to begin capturing attendance.

4. To add a new user, click on the "Ajouter un Nouvel Utilisateur" button and follow the instructions.

Directory Structure

• app.py: Main Flask application file.
• templates/: HTML templates for the web interface.
• static/: Static files including user images and Trained Model.
• haarcascade_frontalface_default.xml: Pre-trained Haar cascade classifier for face detection.

---

Features

1. Face Detection

• Uses a Haar Cascade Classifier (haarcascade_frontalface_default.xml) to detect faces in real time.

2. Face Recognition

• Recognizes faces using a pre-trained KNN model based on data collected from registered users.

3. Attendance Logging

• Attendance is stored in a SQLite database with the following schema:

  Field	Description
  id	Unique record ID
  prenom	Name of the person
  emp_id	Employee ID
  arrivee	Check-in time
  depart	Check-out time
  date	Date of attendance

4. Export Attendance

• Attendance records can be exported as a CSV file for easy reporting.

---

Example Code Snippets

Adding a New User

@app.route("/add", methods=["GET", "POST"])
def add():
    newusername = request.form["newusername"]
    newuserid = request.form["newuserid"]
    userimagefolder = f"static/faces/{newusername}_{newuserid}"
    os.makedirs(userimagefolder, exist_ok=True)
    # Capture images and train the model
    ...
    train_model()

Face Recognition

def identify_face(facearray):
    model = joblib.load("static/face_recognition_model.pkl")
    return model.predict(facearray)

---

Workflow

1. Face Detection: Haar Cascade Classifier detects faces in the webcam feed.
2. Face Recognition: KNN classifier identifies faces based on stored data.
3. Attendance Logging: Recognized individuals are logged into the database with timestamps.
4. Export to CSV: Attendance records can be downloaded.

---

Training the Model

The system uses a KNN model trained on user-provided face images. The training steps include:

1. Extracting face data from images.
2. Resizing and flattening the face images.
3. Splitting data into training and testing sets.
4. Training the KNN model and evaluating metrics.

---

Metrics Visualization

The model evaluation includes:

• Accuracy
• Precision
• Recall
• F1-Score

These metrics are visualized as a bar chart for better understanding.

@app.route("/metrics")
def metrics():
    # Load metrics from a pickle file
    with open("static/metrics.pkl", "rb") as f:
        metrics = joblib.load(f)
    # Generate a bar chart for visualization
    ...

---

Export Attendance

To export attendance as a CSV file:

1. Click the "Export Attendance" button on the web interface.
2. Download the generated CSV file.

def export_to_csv():
    names, rolls, arrivees, departs, _ = extract_attendance()
    csv_output = StringIO()
    csv_writer = csv.writer(csv_output)
    csv_writer.writerow(["Prénom", "N° Emp", "Temps d'arrivée", "Temps de Départ"])
    for name, roll, arrive, depart in zip(names, rolls, arrivees, departs):
        csv_writer.writerow([name, roll, arrive, depart])
    return Response(
        csv_output.getvalue(),
        mimetype="text/csv",
        headers={
            "Content-Disposition": f"attachment;filename=attendance_{datetoday}.csv"
        },
    )

---

Screenshots

Home Page & Attendance

---

Author

scorpionTaj

GitHub Repository: face-recognition