POST

"<base64 encoded image>"

to / to get a list of bounds and confidences like so:

[
  {
    bounds: [x, y, w, h],
    confidence: 0.99
  }
  ...
]

• PORT - the port to listen on
• MIN_CONFIDENCE - minimum confidence to accept
• PADDING - amount (as a ratio) to increase around each detected face (defaults to 0)

• config env vars for blobFromImage params
• sort by area