Face Blurring Tool

A simple and effective Python tool to automatically blur faces in images, videos, or real-time via webcam, using MediaPipe for face detection.

📋 Features

• Image Mode: Blur faces in a static image
• Video Mode: Process a complete video and blur all detected faces
• Webcam Mode: Blur faces in real-time from your webcam

🛠️ Requirements

• Python 3.7+
• OpenCV
• MediaPipe

📦 Installation

1. Clone or download this repository

2. Install the required dependencies:

pip install opencv-python mediapipe

🚀 Usage

Webcam Mode (default)

python face_blur.py --mode webcam

Press ESC to quit.

Image Mode

python face_blur.py --mode image --filePath path/to/image.jpg

The processed image will be saved to ./output/output.png

Video Mode

python face_blur.py --mode video --filePath path/to/video.mp4

The processed video will be saved to ./output/output.mp4

⚙️ Arguments

Argument	Type	Default	Description
--mode	str	webcam	Execution mode: webcam, image, or video
--filePath	str	None	Path to image or video file (required for image/video modes)

📁 Project Structure

.
├── face_blur.py          # Main script
├── README.md             # Documentation
└── output/               # Output folder (created automatically)
    ├── output.png        # Processed image
    └── output.mp4        # Processed video

🔧 Detection Parameters

The script uses MediaPipe Face Detection with the following parameters:

• model_selection: 0 (model optimized for short distance, < 2 meters)
• min_detection_confidence: 0.5 (minimum confidence threshold for detection)

You can adjust these parameters in the code according to your needs.

📝 Examples

# Blur faces in a group photo
python face_blur.py --mode image --filePath photos/group.jpg

# Blur faces in a surveillance video
python face_blur.py --mode video --filePath videos/surveillance.mp4

# Real-time usage
python face_blur.py --mode webcam

🔄 How It Works

1. Face Detection: MediaPipe detects faces in the image/frame
2. Coordinate Extraction: Bounding boxes of faces are retrieved
3. Blur Application: A gaussian blur (30x30) is applied to each detected face
4. Save: The result is displayed (webcam) or saved (image/video)

⚙️ Customization

Modify Blur Intensity

In the process_img() function, change the (30, 30) values:

img[y1:y1 + h, x1:x1 + w, :] = cv2.blur(img[y1:y1 + h, x1:x1 + w, :], (50, 50))

Change Detection Model

# model_selection=0 : short distance (< 2m)
# model_selection=1 : long distance (> 2m)
with mp_face_detection.FaceDetection(model_selection=1, min_detection_confidence=0.5) as face_detection:

Adjust Detection Confidence

# Value between 0.0 and 1.0
# Higher = fewer false positives, but may miss faces
min_detection_confidence=0.7

⚠️ Important Notes

• Default blur is 30x30 pixels
• Output videos use MP4V codec with 25 FPS
• The output/ folder is created automatically
• Output files overwrite previous files

🐛 Troubleshooting

"Cannot open webcam"

• Check that your webcam is connected
• Try changing cv2.VideoCapture(0) to cv2.VideoCapture(1)

"Failed to read video"

• Check that the file path is correct
• Ensure the video format is supported (mp4, avi, mov, etc.)

Slow Performance

• Reduce video/webcam resolution
• Use model_selection=0 for better performance

🤝 Contributing

Contributions are welcome! Feel free to:

• Report bugs
• Suggest new features
• Improve documentation

📄 License

This project is free to use for educational and personal purposes.

🙏 Acknowledgments

• MediaPipe for face detection
• OpenCV for image processing

---

Privacy Note: This tool is designed to protect privacy by blurring faces. Make sure you have the necessary permissions before processing images/videos of people.