Automatic Attendance System (Face Recognition)

Description

• This project is developed in Windows OS.

Working Principle

1. Modules Imported:

• face_recognition: Used for detecting and encoding faces.
• cv2 (OpenCV): For capturing video and displaying the results.
• numpy: For numerical operations.
• csv: To write the attendance logs into a CSV file.
• datetime: To get the current date and time for logging.

2. Camera Initialization:

• The camera is initialized using OpenCV (cv2.VideoCapture(camera)), where the default camera is set to 0. If an external camera is used, you can change this value.

3. Loading and Encoding Known Faces:

• Several face images (e.g., A, Surjith, Ramprasath, etc.) are loaded using face_recognition.load_image_file().
• The faces are encoded using face_recognition.face_encodings() and stored in a list known_face_encoding for later comparison.

4. Known Names:

• The names corresponding to the known face encodings are stored in a list known_faces. These names are used to identify people when their faces are detected.

5. Attendance Logging:

• A CSV file is created with the current date (data/YYYY-MM-DD.csv) to log attendance.
• The program writes the names and the time of detection into the CSV file.

6. Face Detection and Recognition Loop:

• The script continuously captures frames from the camera feed.
• Each frame is resized to 25% of its original size for faster face detection.
• The frame is converted to RGB since face_recognition works with RGB images.
• Face locations and encodings are extracted from the frame.
• Each detected face is compared against the known faces using face_recognition.compare_faces() and face_recognition.face_distance(). The closest match is selected.
• If the face matches a known person, the name is displayed on the video frame, and their attendance is logged.
• If the face is not recognized, it is marked as "Unknown Person".

7. Display and Exit:

• The script displays the video feed in a window labeled "Attendance System". If a known face is detected, it displays the name along with "Present"; otherwise, it shows "Unknown Person".
• The loop continues until the user presses the "q" key, at which point the video capture is released, and the window is closed.

Key Features:

• Face Matching: Recognizes and matches faces in real-time.
• Attendance Logging: Logs the name and the time of appearance in a CSV file.
• Efficient Face Detection: Uses downscaled frames to improve performance without compromising recognition accuracy.

For Installation Process

• Refer the Installation.md file.

Note

• Use the Python 3.10 version.
• Recommended Python Series for the Execution.

Tools

Tools	Version	Check
Python	3.10	[✔]
Pip	latest	[✔]
VSCode	latest	[✔]
PyCharm	latest	[✔]

Show your support

Give a ⭐ if you like this repository!