Recognizing text with cuneiform

There was a task to establish automatic recognition of text from photographs, i.e. the user, when uploading a photograph to the server, also receives the text recognized from it. No sooner said than done. A good free console solution was found - cuneiform . The nix version is here: https://launchpad.net/cuneiform-linux .

So installation. In Ubunt, by the way, version 0.7 is available from the repositories. Version 0.9 is the latest at the moment. The optional argument "-DCMAKE_INSTALL_PREFIX = / your / dir" will install cuneiform in the correct directory. You can run it with the following arguments: Specifies the language of the document. Of the possible ones: eng (default) ger fra rus swe spa ita ruseng ukr srp hrv pol dan por dut cze rum hun bul slo lav lit est tur.

wget http://launchpad.net/cuneiform-linux/0.9/cuneiform-linux-0.9/+download/cuneiform-linux-0.9.0.tar.bz2
tar xvjf cuneiform-linux-0.9.0.tar.bz2
cd cuneiform-linux-0.9.0
mkdir builddir
cd builddir
cmake -DCMAKE_BUILD_TYPE=debug ..
make
make install




-l


-o
Saves to a file.

-f
The format of the received text. Of the supported: text (default), html, rtf, smarttext (plain text with TeX paragraphs), hocr (hOCR HTML format), native (Cuneiform 2000 format)

--dotmatrix
Optimization of the script for an image printed using a matrix printer.

--fax
Optimization of the script for the image printed using fax.

--singlecolumn
Disables page analysis and implies that our image consists of one column of text.

Usage example:
cuneiform -l ruseng -o /our/dir/text.txt /our/dir/book_1.tif

GUI

Then I wanted a graphical interface for everyday needs. There are 2 pieces to choose from - these are YAGF and Cuneiform-Qt :





It was decided to use YAGF. It is also written in qtand requires another spellchecker package aspell. Download, install:

wget http://symmetrica.net/cuneiform-linux/yagf-0.8.1.tar.gz
tar xvfz yagf-0.8.1.tar.gz
cd yagf-0.8.1/
cmake ./
make
make install