You could add some sample results on the github repo (or some tests) with photos of different quality and complexity showing before and face recognised.
Code is mostly taken from the project's github, i just added file selection and image resizing to fit canvas.
Screenshot of Marsellus Wallace from Pulp fiction:
- original: https://vgy.me/7aljyA.png
- contrast lowered: https://vgy.me/fxzO6J.png