Face Recognition Attendance System

Overview

This project implements a face recognition-based attendance system using Python. It captures video from the webcam, recognizes known faces, and marks attendance by recording the recognized names and timestamps in a text file.

Features

• Real-time face recognition using a webcam.
• Attendance marking with timestamp recording.
• GUI interface using Tkinter.
• Saves attendance records in attendance.txt.

Requirements

• Python 3.x
• OpenCV
• face_recognition
• Tkinter (comes with Python)

Installation

1. Install required Python packages:

   pip install opencv-python face_recognition

2. Ensure Tkinter is installed (it usually comes with Python).

Directory Structure

project_directory/
├── E:/MYPIC.png/     # Directory containing known images (JPEG format)
├── attendance.txt    # File where attendance records will be saved
├── attendance_system.py    # Main script file

Usage

1. Place known images: Place the images of known individuals in the E:/MYPIC.png/ directory. The filenames (without extensions) will be used as the names for recognition.

2. Run the script:

   python attendance_system.py

3. The application will open a window showing the video feed from the webcam. Recognized faces will be marked with a green rectangle and their names will be displayed. Unrecognized faces will be marked with a red rectangle.

4. Attendance will be recorded in the attendance.txt file with the format:

   Name,Timestamp
   

Code Explanation

Importing Libraries

The script starts by importing necessary libraries:

import cv2
import face_recognition
import tkinter as tk
import os
from datetime import datetime

Setting Up Tkinter Window

A Tkinter window is created to display the video feed:

window = tk.Tk()
label = tk.Label(window)
label.pack()
button_close = tk.Button(window, text="Close", command=close_window)
button_close.pack()

Loading Known Images

Images of known individuals are loaded, and their face encodings are computed:

known_images_dir = "E:/MYPIC.png"
known_encodings = []
known_names = []
for filename in os.listdir(known_images_dir):
    if filename.endswith(".jpg") or filename.endswith(".jpeg"):
        known_image_path = os.path.join(known_images_dir, filename)
        known_image = face_recognition.load_image_file(known_image_path)
        known_encoding = face_recognition.face_encodings(known_image)[0]
        known_encodings.append(known_encoding)
        known_names.append(filename.split(".")[0])

Initializing Video Capture

The video capture from the webcam is initialized:

video_capture = cv2.VideoCapture(0)

Face Comparison and Attendance Marking

Functions to compare faces and mark attendance are defined:

def compare_faces(known_encodings, face_encoding):
    distances = face_recognition.face_distance(known_encodings, face_encoding)
    threshold = 0.6
    matches = [distance <= threshold for distance in distances]
    return matches

def mark_attendance(name):
    timestamp = datetime.now()
    if name not in marked_names:
        marked_names.append(name)
        with open("attendance.txt", "a") as file:
            file.write(f"{name},{timestamp}\n")

Video Streaming and Face Recognition

The main function to capture video frames, recognize faces, and display the video feed in the Tkinter window:

def video_stream(window):
    ret, frame = video_capture.read()
    face_locations = face_recognition.face_locations(frame)
    face_encodings = face_recognition.face_encodings(frame, face_locations)
    for face_encoding, face_location in zip(face_encodings, face_locations):
        matches = compare_faces(known_encodings, face_encoding)
        if any(matches):
            name = known_names[matches.index(True)]
            mark_attendance(name)
        top, right, bottom, left = face_location
        color = (0, 255, 0) if any(matches) else (0, 0, 255)
        cv2.rectangle(frame, (left, top), (right, bottom), color, 2)
        for match, name in zip(matches, known_names):
            if match:
                cv2.putText(frame, name, (left, top - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, color, 2)
    frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
    img = tk.PhotoImage(data=cv2.imencode('.png', frame_rgb)[1].tobytes())
    label.config(image=img)
    label.image = img
    if window.winfo_exists():
        window.after(1, video_stream, window)

Starting the Video Stream

The video stream is started and the Tkinter event loop is run:

video_stream(window)
window.mainloop()
video_capture.release()

Notes

• Adjust the video_capture index if using a different camera.
• Modify the threshold value in the compare_faces function if needed to adjust recognition accuracy.

License

This project is licensed under the MIT License.