[
https://issues.apache.org/jira/browse/TIKA-4946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18125203#comment-18125203
]
Tim Allison commented on TIKA-4946:
-----------------------------------
I have a fairly robust and simple calculation that works for this. This is
worth pursuing. We'll still miss pages with some noise on them and send them to
OCR when they don't even contain text, but we can at least rule out largely
completely blank pages without false positives.
> Don't send blank pages for OCR
> ------------------------------
>
> Key: TIKA-4946
> URL: https://issues.apache.org/jira/browse/TIKA-4946
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Major
>
> Some vlm-based OCR tools will hallucinate an entire journal article page's
> worth of fluent content if we send a blank image.
> Ideally, the ocr tools would prevent this, but we should try to do something
> on our side to prevent this.
> Not sure where to put this: in the PDFParser or in the OCR engines or both?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
