Note: I am not sure if I understand this correctly. I have not used the app yet. Please correct me if my angle is all wrong!
In step 2 of your getting started guide, there is considerable manual effort.
Suppose that instead you have some glyph segmentation from an external annotation format and workflow, like OCR output in hOCR or PAGE-XML or ALTO-XML format. (Really, any suitable segmentation-providing format would do.)
Couldn't Glyphcollector then use these coordinates to crop sample images, perhaps with a heuristic to keep some subset of visually maximally distant or textually (i.e. predefined) samples and either ignore the others or use them for bootstrapping your search?
Also, would external high-quality binarization (i.e. bitonal reduction) help in any way?
In step 2 of your getting started guide, there is considerable manual effort.
Suppose that instead you have some glyph segmentation from an external annotation format and workflow, like OCR output in hOCR or PAGE-XML or ALTO-XML format. (Really, any suitable segmentation-providing format would do.)
Couldn't Glyphcollector then use these coordinates to crop sample images, perhaps with a heuristic to keep some subset of visually maximally distant or textually (i.e. predefined) samples and either ignore the others or use them for bootstrapping your search?
Also, would external high-quality binarization (i.e. bitonal reduction) help in any way?