Opened 4 years ago

Last modified 2 years ago

#304 new defect

Cannot search searchable DjVu files with OCR layer

Reported by: Lewis Rosenthal Owned by:
Priority: minor Milestone: 1.4.2
Component: Plugin: DjVu Version: 1.3.6
Keywords: Cc:


Testing has revealed that we are either unable to render the OCR layer of "searchable" DjVu files or we do not recognize them as searchable. We're in good company, though, as I was not able to get Okular on Linux to search them, either. Instead, both apps just see them as image files.

For my test, I converted a true PDF (generated text) to DjVu using pdf2djvu (on Linux), and then ran ocrodjvu against that, which processes the text using tesseract. I find it hard to believe that DjVuLibre cannot recognize the OCR layer, yet both Lucide and Okular saw no apparent difference between the two versions.

We should consider investigating this, if there is sufficient interest in searching DjVu files.

Change History (1)

comment:1 Changed 2 years ago by Gregg Young

Milestone: Future1.4.2
Note: See TracTickets for help on using tickets.