Noise based on Face Recognition

[ThoughtfulDev]

This function automatically noises a person in an Image based on one (or more) Images of that Person.
I have added new CLI Arguments such as:

  -f, --facerecog       Does face recognition
  -k KNOWN_PATH, --known KNOWN_PATH
                        Path to known images

They should be fairly simple.
Example call would be
python person_blocker.py -i block.jpg -f -k ./known where 'known' is a folder with jpgs of the person you want to block.

This pull request adds the following dependencies

• opencv-python
• face_recognition (which itself has dlib as a dependency)

dlib can be installed on Windows but its a pain in the a**

[minimaxir]

This adds a lot of complexity to the script (and opencv-python is not an easy dependency to add either).

I'll have to think about it.

[ThoughtfulDev]

OpenCV has only been used on the resize part (which could easily be skipped). So CV isn't really necessary.

Edit: I removed the OpenCV dependency.

I just wanted to make it as close to Blackmirror as possible.