I'm working on setting up a Quickfields session to redact account numbers, social security numbers, and the like. I'm trying to have the redactions leave the last 4 digits of each number unredacted. As an example, my regular expression for finding and redacting Social Security numbers is:

(\d\d\d[ -.|]\d\d[ |.-])\d\d\d\d(?x)

Some of the time, it works great. Other times, it redacts all of the digits in the image, or redacts stuff before the SSN in the image. (The attached png file shows a case where the word before the social security number is redacted in the image, but the text redaction is correct.) The redaction in the text is always correct (just the first digits redacted with the last 4 unredacted). I also have one case where the document is failing to save to the repository once it's finished because it's trying to create a redaction rectangle that's bigger than the page for some reason.

I think all of these problems are related: it seems like my image/text link is a little messed up so Quickfields is unsure of where the OCR text actually lies in the image. Does anyone have any ideas for what I can try to do to troubleshoot this?