OCR with contextΒΆ

OCR should use context, when it sees the word ‘everythxxg’, it should know that the ‘xx’ is ‘in’. This is how the human brain works, and is how the computer should work too. If Google can suggest spelling suggestions to my misspelled words, there is no reason that this technology couldn’t be applied to OCR. It would make it much more powerful and useful.



Hey there! I'm Eric and I work on communities in the world of software documentation. Feel free to email me if you have comments on this post!