sbuss / ocropy Goto Github PK

To install, use:

    $ sudo apt-get install $(cat PACKAGES)
    $ python setup.py download_models
    $ sudo python setup.py install

To test the recognizer, run:

    $ ./run-test

OCRopus is really a collection of document analysis programs, not a turn-key OCR system.

In addition to the recognition scripts themselves, there are a number of scripts for
ground truth editing and correction, measuring error rates, determining confusion matrices, etc.
OCRopus commands will generally print a stack trace along with an error message;
this is not generally indicative of a problem (in a future release, we'll suppress the stack
trace by default since it seems to confuse too many users).

To recognize pages of text, you need to run separate commands: binarization, page layout
analysis, and text line recognition. 

    # perform binarization
    ./ocoropus-nlbin tests/ersch.png -o book

    # perform page layout analysis
    ./ocropus-gpageseg 'book/????.bin.png'

    # perform text line recognition (on four cores, with a fraktur model)
    ./ocropus-rpred -Q 4 -m models/fraktur.pyrnn.gz 'book/????/??????.bin.png'

    # generate HTML output
    ./ocropus-hocr 'book/????.bin.png' -o ersch.html

    # display the output
    firefox ersch.html

There are also a number of older commands for text line recognition,
layout analysis, etc., kept for backwards compatibility. The binarization
and layout analysis commands of the current release will be replaced in 
the next release with entirely new, trainable commands.

The main feature of this release is ocropus-rpred, which achieves very
low error rates on a wide variety of fonts and inputs (even degraded)
of body text, even without language models or dictionaries. The model
has been trained on UW3 and UNLV data.  Test set error on UW3 is
about 0.5% without a language model or dictionary.

There are some things the currently trained models for ocropus-rpred
will not handle well, largely because they are nearly absent in the
current training data. That includes all-caps text, some special symbols
(including "?"), typewriter fonts, and subscripts/superscripts. This will
be addressed in a future release, and, of course, you are welcome to contribute
new, trained models.