Alphabot - a screen-less interactive spelling primer powered by computer
vision. This program comes with ABSOLUTELY NO WARRANTY. This is free software,
and you are welcome to redistribute it under certain conditions.

The suggested alphabet letter set is "Melissa & Doug 52 Wooden Alphabet Magnets in a Box",
available from Amazon at https://www.amazon.ca/Melissa-Doug-Wooden-Alphabet-Magnets/dp/B001990BUU/.

Please see https://medium.com/@drewgillson/alphabot-a-screen-less-interactive-spelling-primer-powered-by-computer-vision-cc1095bce90a

Copyright (C) 2018 Drew Gillson <drew.gillson@gmail.com>

Usage: python3 alphabot.py

optional arguments:
  --input               OpenCV device id for your top-mounted camera, default is 0
  --width               Video width, default is 800
  --height              Video height, default is 600
  --length              Default starting character length for words, default is 3
  --debug [True|False]  Print debug information and display input to the letter
                        detector CNN using imshow, default is True
  --corpus              The CSV file in corpus/ containing words you will be
                        prompted to spell, default is One Fish Two Fish Red Fish
                        Blue Fish, Seuss, Dr. (1960).
  --random [True|False] Randomize word selection from corpus
  --tts                 Text-to-speech binary, default is Mac OS "say" command
  --train               Re-train the letter detection CNN by using the training data in images/