Behavior:
Intelligent Indexing does not recognize all areas of the document, so some index terms have to be entered manually.
Solution:
In order for Intelligent Indexing to recognize all configured index terms and fill the index fields with them, the relevant areas on the document must be machine-readable and have sufficient quality.
If certain areas of a document cannot be read by Intelligent Indexing, there is no text shot for these areas in the original document (PDF).
In order to make these areas usable for Intelligent Indexing, the OCR must generate a new text shot for the document so that the complete document is made available for the full text and for Intelligent Indexing.
In cloud systems, this can be implemented via an import job. To force the creation of a new text shot during the import job, only one file must be adapted.
In the file "...\DocuWare\Desktop\DocuWare.DesktopService.exe.config", add the following key to the <appSettings> section:
<add key="UseOcrForNativePdf" value="true"/>.
Then restart the DocuWare Desktop Service.
For On Premises systems you can force the creation of a new text shot in general, e.g. also when dragging and dropping into the mailbox.
To do this, add/change the following value in the "DocuWare.Imaging.Worker.exe.config" file in the \DocuWare\Common\Imaging\ directory.
Previously:
<TextExtractionMethods>.
<add fileType="Raster" method="Ocr"/>
</TextExtractionMethods>
After:
<TextExtractionMethods>
<add fileType="Raster" method="Ocr"/>
<add fileType="Pdf" method="ToOcr"/>
</TextExtractionMethods>
Both changes have the effect that the processing time of the documents increases slightly, because the OCR needs time to create the new text shot.
