查看“︁OCR”︁的源代码

= Tesseract OCR =
== Build tesseract ==

Prerequisits<ref>https://stackoverflow.com/questions/33659458/tesseract-image-issue</ref>:

<syntaxhighlight lang="bash">
brew install libgif libjpeg libpng libtiff zlib

# Error: xz: undefined method `deny_network_access!' for Formulary::FormulaNamespaceeddce1918855a2fb5cf7427fd5438072::Xz:Class
# then comment this line `# deny_network_access! [:build, :postinstall]`
</syntaxhighlight>


Install leptonica<ref>https://github.com/DanBloomberg/leptonica</ref>:

<syntaxhighlight lang="bash">
# https://stackoverflow.com/questions/40067547/glibtool-on-macbook
brew install libtool automake

git clone https://github.com/DanBloomberg/leptonica.git
cd leptonica
./autogen.sh
./configure 
make 
sudo make install
</syntaxhighlight>

Install tesseract-ocr:

<syntaxhighlight lang="bash">
git clone https://github.com/tesseract-ocr/tesseract.git
cd tesseract
./autogen.sh
./configure
make
sudo make install
</syntaxhighlight>

== Trained data ==

We have three sets of official .traineddata files trained at Google, for tesseract versions 4.00 and above. These are made available in three separate repositories.

* tessdata_fast (Sep 2017) best “value for money” in speed vs accuracy, Integer models.
* tessdata_best (Sep 2017) best results on Google’s eval data, slower, Float models. These are the only models that can be used as base for finetune training.
* tessdata (Nov 2016 and Sep 2017) These have legacy tesseract models from 2016. The LSTM models have been updated with Integer version of tessdata_best LSTM models. (Cube based legacy tesseract models for Hindi, Arabic etc. have been deleted).

<syntaxhighlight lang="bash">
git clone https://github.com/tesseract-ocr/tessdata_best.git

</syntaxhighlight>
[[Category:Deep Learning]]