PATCH: preserve interword spaces

[dhara...@gmail.com]

I am attaching a small patch that adds a configurable parameter that when enabled will preserve multiple interword spaces in text.  The strategy for doing this was based on Ray's suggestion from several years ago in [1]. This seems to be useful to others as well [2] [3] [4].

As Ray mentioned originally, the number of spaces output will vary/be inaccurate.  That's ok for our use case, as we are really more interested in detecting one space vs. multiple spaces.

I am happy to make adjustments if necessary.

Thanks,

David

[1] https://groups.google.com/forum/#!searchin/tesseract-ocr/spaces/tesseract-ocr/lGBQiryHcrY/wy5a-L9O3i4J

[2] https://groups.google.com/forum/#!searchin/tesseract-ocr/spaces/tesseract-ocr/9nzPrBZ3118/b3W5GtsFPo0J

[3] http://stackoverflow.com/questions/22609778/how-to-preserve-document-structure-in-tesseract

[4] http://stackoverflow.com/questions/18089013/tesseract-ambiguity-in-space-and-tab

[zdenko podobny]

Hi,

thanks for patch. Can you please provide also some test case/image that would demonstrate problem?

Zdenko

--
You received this message because you are subscribed to the Google Groups "tesseract-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-de...@googlegroups.com.
To post to this group, send email to tesser...@googlegroups.com.
Visit this group at http://groups.google.com/group/tesseract-dev.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-dev/5c5bdde5-4886-469f-a5af-b8097b0b5587%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[dhara...@gmail.com]

Here is a sample image, with differing outputs...

tesseract -l eng -c preserve_interword_spaces=1 109359.tiff output-preserve-enabled

tesseract -l eng 109359.tiff output-default

Thanks.

David

[zdenko podobny]

Thanks. Committed in Revision: 36883b4fafcd.

Zdenko

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-dev/1c93c4a3-ad14-49b0-a461-f5f7c84fc6b9%40googlegroups.com.

[Jan Ruzicka]

Hi,
There is one small problem with the patch.
Now there is a space at the beginning of each line in case of preserve_interword_spaces_ == false.
Previously the space was added after the word and last one was removed.

Adding (words_appended > 0) instead of the 1 in the numSpaces calculation would work, but it looks ugly.

so instead of:
int numSpaces = preserve_interword_spaces_ ? it_->word()->word->space() : 1;

it would be:
int numSpaces = preserve_interword_spaces_ ? it_->word()->word->space() : (words_appended > 0);

Jan

> To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-dev/CAJbzG8xQMgUqgDDy9bQDmoKS-Lh8CYomaS2gfPG1zKgKn-g_Qw%40mail.gmail.com.

[dhara...@gmail.com]

Oops, my apologies.  Thank you Jan.  I suppose I was too focused on the output with the parameter enabled :-)

Since the fix is so minor it is perhaps easiest if maybe Zdenko just changes it directly?

resultiterator.cpp:648

David

[zdenko podobny]

committed.

Zdenko

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-dev/55a57199-bb12-482f-9628-f5e36ad2fbee%40googlegroups.com.

[Dmitri Silaev]

FYI. This patch causes a compilation error under MSVC 2010 and earlier. And from what I know, same in later versions (not tested). This is due to C++11 member initialization inside class declaration:

File: "ccmain\resultiterator.h"
Ln 239: bool preserve_interword_spaces_ = false;

Best regards,
Dmitri Silaev

www.CustomOCR.com

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-dev/CAJbzG8yynM5F6GadLcKGs-VkuRa8GW6%3Dd0vkwmswuAPHKMGhRA%40mail.gmail.com.

[dhara...@gmail.com]

The initialization can be moved to the ResultIterator constructor... ccmain/resultiterator.cpp:37 

Thanks Dmitri.

David