Perfect for profile picture processing for your website or batch work for ID cards, autocrop will output images centered around the biggest face detected.


From the command line:

usage: [-h] [-o OUTPUT] [-i INPUT] [-w WIDTH] [-H HEIGHT] [-v]

Automatically crops faces from batches of pictures

optional arguments:
  -h, --help            Show this help message and exit
  -o, --output, -p, --path
			Folder where cropped images will be placed.
			Default: current working directory
  -i, --input
			Folder where images to crop are located.
			Default: current working directory
  -w, --width
			Width of cropped files in px. Default=500
  -H, --height
			Height of cropped files in px. Default=500
  --facePercent   Percentage of Face height to image height (zoom factor)
  --padUp         Padding up value compared to padDown. Default=50
  --padDown       Padding down value compared to padDown. Default=50
  --padLeft       Padding left value compared to padRight. Default=50
  --padRight      Padding right value compared to padLeft. Default=50
  -v, --version         Show program's version number and exit

Params (width, height, facePercent)

  • Example:
    autocrop -i pics -o crop -w 400 -H 400 --facePercent 50.
  • Example more padding down:
    autocrop -i pics -o crop -w 400 -H 400 --facePercent 50 --padUp 20 --padDown 50.

What it does

The previous command will:

  1. Copy all images found in the top level of pics to crop,
  2. Crop around the face and resize to 400x400 pixels all images in crop.

Images where a face can't be detected will be left in crop.
If no output folder is added, asks for confirmation and destructively crops images in-place.


Simple! In your command line, type:

pip install autocrop


Autocrop uses OpenCV to perform face detection, which is installed through binary wheels. If you already have OpenCV 3+ installed, you may wish to uninstall the additional OpenCV installation: pip uninstall opencv-python.

Installing directly

In some cases, you may wish the package directly, instead of through PyPI:

cd ~
git clone
cd autocrop
pip install .


Development of a conda-forge package for the Anaconda Python distribution is also currently slated for development. Please leave feedback on issue #7 if you are insterested in helping out.


Best practice for your projects is of course to use virtual environments. At the very least, you will need to have pip installed.

Autocrop is currently being tested on:

  • Python:
    • 2.7
    • 3.4
    • 3.5
    • 3.6
  • OS:
    • Linux
    • macOS
    • Windows