🎥 Real-Time Face Blur using OpenCV DNN

A real-time computer vision project that detects human faces from a live webcam feed and automatically blurs them using OpenCV's Deep Neural Network (DNN) module. This project demonstrates privacy-preserving video processing using a pre-trained SSD face detection model.

---

✨ Features

• 🎯 Real-time face detection
• 👤 Automatic Gaussian blur for detected faces
• 📹 Live webcam video processing
• ⚡ Fast inference using OpenCV DNN
• 🚫 Displays "No Face Detected!" when no faces are found
• 🧠 Uses a pre-trained SSD Caffe face detection model

---

🛠️ Technologies Used

• Python
• OpenCV
• NumPy
• OpenCV DNN Module
• Jupyter Notebook

---

📂 Project Structure

Face_Blur_Realtime/
│
├── main1.ipynb
├── deploy.prototxt
├── res10_300x300_ssd_iter_140000_fp16.caffemodel
README.md
requirements.txt

---

🧠 Pre-trained Model

The project uses OpenCV's pre-trained SSD face detector.

Model files:

• deploy.prototxt
• res10_300x300_ssd_iter_140000_fp16.caffemodel

---

🚀 Installation

Clone the repository

git clone https://github.com/Pheonix-1002/Realtime_FaceBlur.git

cd Realtime_FaceBlur/Face_Blur_Realtime

Install dependencies

pip install opencv-python numpy jupyter

---

▶️ Run the Project

Start Jupyter Notebook:

jupyter notebook

Open:

main1.ipynb

Run all cells to start the webcam.

Press q to exit the application.

---

⚙️ Workflow

1. Capture frames from the webcam.
2. Convert each frame into a blob.
3. Perform face detection using OpenCV's DNN model.
4. Extract the detected face region.
5. Apply Gaussian Blur.
6. Display the processed video stream.

---

📈 Future Improvements

• Face pixelation mode
• Blur intensity adjustment
• Video file support
• Face tracking
• YOLO-based face detection
• GPU acceleration using CUDA

---

👨‍💻 Author

Pheonix-1002