Face Detection Project

A Python-based face detection system using OpenCV and Haar Cascades for real-time face detection and dataset creation.

Features

• Real-time face detection using webcam
• Automatic face dataset creation
• Face cropping and resizing
• Image storage in organized directories

Requirements

• Python 3.x
• OpenCV (cv2)
• Webcam/Camera access

Installation

1. Install required dependencies:

pip install opencv-python

2. Ensure your camera/webcam is connected and accessible

Project Structure

face-detection/
├── Day5_Code/
│   ├── dataset/
│   │   └── champ/          # Captured face images
│   ├── FaceDetection.py    # Main face detection script
│   ├── faceDataset.py      # Dataset creation script
│   └── haarcascade_frontalface_default.xml  # Haar cascade classifier
└── README.md

Usage

Face Detection with Dataset Creation

Run the main face detection script:

python Day5_Code/FaceDetection.py

Alternative Dataset Creation

Run the dataset creation script:

python Day5_Code/faceDataset.py

How It Works

1. Camera Initialization: Opens webcam feed
2. Face Detection: Uses Haar Cascade classifier to detect faces
3. Image Processing: Converts to grayscale and crops detected faces
4. Storage: Saves resized face images (130x100 pixels) to dataset folder
5. Real-time Display: Shows live feed with face detection rectangles

Controls

• ESC key: Exit the application
• Automatic capture: Captures 30 face images automatically

Output

• Face images are saved in dataset/champ/ directory
• Images are numbered sequentially (1.jpg to 30.jpg)
• Each image is resized to 130x100 pixels in grayscale

Troubleshooting

• Camera not opening: Check camera permissions and ensure no other application is using the camera
• No faces detected: Ensure proper lighting and face positioning
• Import errors: Verify OpenCV installation with pip install opencv-python

Notes

• The system captures 30 images by default
• Face detection works best with good lighting conditions
• Ensure face is clearly visible and properly positioned in the camera frame