* convert into dict output
* faces class
* return cleaned dict
* empty methods that are required
* with dict updates
* with dominant emotion confidence as tuple
* multiple images
* update notebook
* read image into nb
* test
* added keras-ocr and google vision
* google cloud vision by far the best
* setting up docker for text 1
* move widgets and analysis to display module
* move widgets and analysis to display module - 2
* text on image through widgets