Suggestion to improve OCR

Hey there, just ran a number of documents through the engine and noticed that there were a few instances where a currency (say, $650.23), was recognized as $65o.23 (it recognized ‘o’ instead of zero). Not sure if you have any embedded contextual logic in your engine, but I’d suggest that if the engine detects a letter (at least an ‘I’ or ‘o’) in the middle of a bunch of numbers, especially numbers preceded by a $, it bias the result towards a number rather than a letter? Is that possible?

Hi, are you using engine1 or 2? Also, can you please post some example images?

I’m using engine #2. I wrote some code to fix it in post-processing, but will add something to let me know when it triggers and will send over an example or two next time I see it.

Hey folks, I’ve upgraded from PDF to PDF Pro. Is there some sort of refund available? Thanks.