What is Force OCR?
When processing PDFs, Parseur can extract data in two ways:
Embedded Text Extraction: If a PDF contains embedded text, Parseur will extract data directly from this text.
Optical Character Recognition (OCR): If a PDF does not contain embedded text, Parseur uses OCR to detect and extract text visually from the document.
In most cases, extracting data from embedded text is more accurate than using OCR. OCR relies on computer vision and machine learning models, which, while powerful, are not always 100% accurate. Parseur uses advanced OCR technology with excellent recognition capabilities, but the results can still vary.
Why Use Force OCR?
Sometimes, the embedded text in a PDF is incorrect or garbled, often due to errors in the software that created the embedded text, especially with poorly configured OCR in scanned documents.
In such cases, enabling the Force OCR option instructs Parseur to always use OCR to extract text, even if the PDF already contains embedded text. However, be aware that forcing OCR can reduce the quality of the extracted text if the embedded text is already accurate. Therefore, use the Force OCR option with caution.
How do I enable Force OCR?
To turn on the Force OCR option:
Open your Parseur mailbox.
Click on Settings in the left menu.
Navigate to the Processing tab.
Under Advanced Settings, check the box labeled Force use of OCR on PDFs.
Click Save.
Once enabled, this setting will apply to any new PDFs uploaded. To apply it to an existing document, you need to reprocess that document.
When should I use Force OCR?
Consider using the Force OCR option in the following scenarios:
The text extracted from your PDFs is incorrect or appears as gibberish (e.g., random special characters).
Some text is missing from the extracted data.
You encounter a “Label not found” error when creating labels in an OCR template for certain text.
Note: The Force OCR setting applies to all documents within a mailbox. If you work with a mix of documents—some needing OCR and some not—it is best to create two separate mailboxes: one with Force OCR enabled and another without. Upload documents to their respective mailboxes accordingly.