Face Detection and Censoring System

A Python-based system for detecting faces in images and videos using YOLOv8, with the ability to censor detected faces. The system is designed to be modular and extensible.

Features

• Face detection using YOLOv8
• Support for both image and video processing
• Easy to use User Interface
• Modular censoring system
• Trained on the WIDER FACE dataset via Roboflow
• Multiple masking methods including: blur, emoji, and text (see demo)

Local Installation

1. Clone the repository:

git clone https://github.com/Spring-0/face-censor.git
cd face-censor

2. Create a virtual environment and activate it:

python -m venv .venv
source .venv/bin/activate  # On Windows, use: .venv\Scripts\activate

3. Install the required packages:

pip install -r requirements.txt

4. Run:

python src/main.py

Training the Model - Optional

The project uses the WIDER FACE dataset from Roboflow for training. I have included a pre-trained model, so there is no need to re-train it unless you want to. Here is how:

1. Update this line in training/training.py if required:

device="0"  # Set to "0" to utilize GPU, otherwise set to "cpu" to utilize CPU

2. Create a .env file in the project root with your Roboflow API key:

ROBOFLOW_API_KEY=your_api_key_here

3. Run the training script:

cd training
python3 training.py

Usage

API Usage

# Face detection model
from models.yolo_detector import YOLOFaceDetector

# Masking methods (no need to import all, just what you want to use)
from masking.text import TextCensor
from masking.emoji import EmojiCensor
from masking.blur import BlurCensor

# Media processor
from processor import MediaProcessor

# Initialize face detector model
detector = YOLOFaceDetector()

Creating Masking Object

This is what determines what effect will be applied to mask the faces.

Using Text Masking

text_censor = TextCensor(
    text="HELLO", # The text to draw on faces
    draw_background=True, # Control whether to draw solid background behind text
    background_color="white", # The color of the solid background
    text_color="black", # The color of the text
    scale_factor=0.2 # The text size scaling factor, default to 0.5
)

Using Emoji Masking

emoji_censor = EmojiCensor(
    emoji="😁", # The emoji you want to use to mask faces
    font="seguiemj.ttf", # The path to the emoji font file, by default uses "seguiemj.ttf"
    scale_factor=1.0 # The emoji size scaling factor in percentage, default to 1.0
)

Using Blur Masking

blur_censor = BlurCensor(
    blur_factor=71 # The strength of the blur effect, defaults to 99
)

Create Media Processor

After creating the masking method object(s), you need to pass it to the MediaProcessor constructor like so:

processor = MediaProcessor(detector, blur_censor)

Processing Images

# Process an image
processor.process_image("input.jpg", "output.jpg")

Processing Videos

# Process a video
processor.process_video("input.mp4", "output.mp4")

Demo

Input Image/Video

Output Image/Video

Blur Masking

Emoji Masking

Text Masking

Requirements

• Python 3.8+
• PyTorch
• OpenCV
• Ultralytics YOLOv8
• Roboflow

See requirements.txt for complete list.

License

GPU General Public License - see LICENSE file for details.

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

TODO

• Add emoji face masking
• Add support for real time streams
• Add GUI interface
• Add partial face censoring (eyes)