I don't see why there can't be.If you can OCR some text, by matching the skeleton of each letter, then each font would have letter backbones with different deviations from some ideal.
Or, at least, the difference between the sample font letters and each of the target fonts would be a measure which could be used to rank the candidate matches.
There might be some speedups to the search if certain letters were more discriminating than others, perhaps font-dependant, so that letters would be compared in order of discriminating ability.
(Or, at least, the difference between the sample font letters and each of the target fonts would be a measure which could be used to rank the candidate matches.
There might be some speedups to the search if certain letters were more discriminating than others, perhaps font-dependant, so that letters would be compared in order of discriminating ability.
The Alchemist king of the needlessly complicated,
Wed 24 Feb 2010, 14:03,
archived)