Unfortunately, the answer here is "it's complicated".

It's really hard to say when it the problem in general will be fixed, as OCR technology always involves a bit of high-tech guesswork and thus is likely never going be 100% accurate for every image. Since Laserfiche uses a 3rd party OCR engine we don't always have control over the inner workings of the OCR.

However, the good news is we ARE looking into various approaches to improving OCR accuracy over the next year, so it's still possible you will see improvements here. If you're interested in sending us specific examples of "problematic" images that we can use to help judge accuracy in a way that better meets your needs, please open a support case, attach the relevant image in its original format, and ask them to send it to me and/or the capture team.