word/glyph segmentation

[bertsky]

Is there any chance we can get the horizontal positions associated with the decoded characters? We normally use this to map them back to Word and Glyph coordinates (see ocrd_calamari, ocrd_kraken, ocrd_ocropy for comparison).

[GemCarr]

We have no way of getting these positions easily atm. However it is not out of the question for future improvements, as we have some ideas that could work for implementing this.