Does the MyScript SDK allow for creation of custom recognition databases?
I have been approached to research the possibility of using handwriting recognition software to recognize Gregg and/or Pitman shorthand, and I'm willing to do the work of compiling the stroke data, the vocabularies, and corpus training materials, but I don't know if the MyScript SDK allows sufficiently low-level access to the recognition engine itself.
I realize that there are some extremely tricky issues in recognizing shorthand, but I'm willing to entertain the possibility of adding my own pre- and post-processing to the incoming stroke data, and to the textual recognition results, in order to achieve higher accuracy. The pen-strokes-to-text pattern recognition engine is the wheel which I do not wish to re-invent.
Did you try one of our widget or our cloud offer? Specific resources can be created such as lexicon to adapt the text recognition, however it won't address your use case. Each language pack also includes default resources: alphabet knowledge to define the characters and symbols (built from our ink database), linguistic knowledge and grammar (built from our corpora). Addressing your needs would require to create a new specific language and its corresponding resources: it cannot be done on your side.
The only way to address such use case would be to consider creating a dedicated R&D project, hence understanding more your budget and schedule. If you are interested, please fill in this form.