Open ion-oset opened 3 years ago
I prefer not to use a barcode, and to have the text used for identification to be human comprehensive rather than (for example) an opaque string of chars that could encode anything.
The third option I am not sure I understand, because I don't know where an image comes in (except for the conventional placement of an image of state or county seal); I don't think that anyone has done that before; for that matter I don't think it has been done based human readable+comprehensible text at all
Progress: can now extract image pixels of page headers; next up is aligning and matching a new unknown page to find the header that fits best. https://github.com/brianolson/ballotstudio/commit/39661f344ab95164a09c58ec0d45039c27c7faaa
this works in a development tool imdev
as of 7fd88785d94409dd83840a5072de312f298a9d2a
Add identification marks to a ballot so that a scanner configured broadly for "an election" can scan any ballot style in that election. This can take several forms. Examples