This simple program was developed to help in generation datasets that you can
use for facial detection and recognition models.
The program uses OpenCV's haarcascade_frontalface_alt2.xml for
face detection then the detected face is saved in the folder you specify.
For this program images would be saved under a folder called images already
specified for you and a subfolder which you'll specifiy yourself, e.g martin
- Create a virtual environment:
virtualenv dataset
- Start the environment: (Linux)
source dataset/bin/activate
- Use pip to install dependencies:
pip3 install -r requirements.txt
- Run the generator:
python3 dataset_maker.py
This should prompt you for a folder name as below
Once you click Entr, a video stream will start showing. A green bounding box indicates that images are being saved. - The directories are
- Sample image
![image6] - To quit press Q