Correcting OCR inaccuracies

I have Hazel rules to act on my documents after they are scanned and OCR’d (by PDF pen). Sometimes the scanning is less than accurate such that one of the criteria Hazel needs becomes nonsense (December 8, 2014 becomes December Q 2014) and all the downstream Hazel actions are lost.

I tried using my PDFpen to put the correct text on the document hoping Hazel can do her thing, but to no avail. Is there a way to add text to a document in a way that Hazel can see it and therefore continue on with her instructions?

Thanks in advance.

I don't know of a way off-hand but maybe you should contact the PDFPen developers.

Although it is pretty expensive ($120), I have found that ABBYY FineReader OCR Pro has the best accuracy of all the various OCR software products. The type of error you refer to doesn't happen insofar as I have noticed with this product, as I think it takes the context of the OCR process into account (i.e., it's smart enough to know that what follows "August" is much more like to be an "8" instead of a "Q"). ABBYY added automator actions to a recent release and I now use this software for all my Hazel OCR activities.

I contacted PDFPen and got a fix. They point out that OCR is never perfect, no matter which product you use. ABBY Finereader may be more accurate, maybe not, but I believe manual corrections will be needed occasionally with any software.

For the PDFPen:
Go to View -> OCR layer, then find the mistaken recognition, double click on it, click "correct text" in top bar or under the "format" menu, then type the change.
This will only change the searchable text; it won't change the actual text visible on the document.

I tried this and Hazel recognized the change, and was able to act on it appropriately. I am overjoyed.