BIP😸 is an informational BIP that outlines a general guideline for encoding Bitcoin addresses as strings of emoji. It includes example code for working with emoji addresses.
-
Grab all images from here: https://unicode.org/emoji/charts/full-emoji-list.html
-
Parse the table using this: https://stackoverflow.com/questions/6325216/parse-html-table-to-python-list
-
Save each image as a PNG with a name like: cldr_short_name-platform.png
Note:
00:47 < Terminus> faceface: echo
"R0lGODlhEAAMAKIFAF5LAP/zxAAAANyuAP/gaP///wAAAAAAACH5BAEAAAUALAAAAAAQAAwAAAMsWLPcrSHKOVcgOOtsiRCbBxLdF5rkcIVhx27uy6kagNlafOM5LXMKh3BRSAAAOw=="
| base64 -d > ~/foo.gif 00:47 < Terminus> it's a gif btw so the mimetype is wrong.
-
Apply random transformations using Keras: https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html
-
Try to classify cldr regardless of platform...
Note: Find the list of cldr that are most accurately classified across platforms (e.g. those that are most similar / least confused across platforms).
Note: Classification or OCR?
-
Create a list of 58 robustly classified emoji across platforms.
-
Map them to base 58 characters... Somehow.
-
Create libraries to work with this encoding.