Dear all, I have been testing tesseract to embed OCR in scanned PDF documents, and it works phenomenally well in recognizing the text.
Now I noticed one slightly disturbing issue just by chance when comparing the original input image and the PDF file: A number of straight lines that are present in the input image have disappeared completely in the PDF (some of the are horizontal rules, others are lines in a logo). Since I wanted to use tesseract to produce completely unmodified documents with only the OCR text layer added, this would be a problem for me. I have uploaded a test image for this to http://cern.ch/fsiegert/tmp/tesseract-test.tif and here is the command I used on it: > $ tesseract -l deu tesseract-test.tif tesseract-test pdf > Tesseract Open Source OCR Engine v3.03 with Leptonica > OSD: Weak margin (6.96) for 162 blob text block, but using orientation > anyway: 1 > $ tesseract --version > tesseract 3.03 > leptonica-1.71 > libgif 5.1.0 : libjpeg 8d : libpng 1.6.12 : libtiff 4.0.3 : zlib 1.2.8 : > libwebp 0.4.1 This results in http://cern.ch/fsiegert/tmp/tesseract-test.pdf, which is missing the straight horizontal lines and the ones in the logo. Is this line-removal done on purpose and can it be disabled? Cheers, Frank PS: I have removed much more text from the document for privacy reasons, but the same happens when the document is complete with text. -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/9d3455ba-6c17-4c10-bc09-e5ee5b911ad0%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
