Use ocrd_olena for binarization

This commit is contained in:
Gerber, Mike 2019-10-21 17:04:06 +02:00
parent 63c364207c
commit 7025d960b4
2 changed files with 16 additions and 4 deletions

View file

@ -15,11 +15,13 @@ RUN apt-get update && \
python3-pip \
git \
# For leptonica/tesseract:
cmake libgif-dev libjpeg-dev libpng-dev libtiff-dev zlib1g-dev \
cmake libgif-dev libjpeg-dev libpng-dev libtiff-dev zlib1g-dev libpango1.0-dev \
# For clstm on Ubuntu 19.04:
swig libeigen3-dev libpng-dev libprotobuf-dev \
# For cv2:
libsm6 libxrender1 \
# For ocrd_olena:
wget graphviz imagemagick libmagick++-dev libgraphicsmagick++1-dev libboost-dev \
# XML utils
libxml2-utils \
xmlstarlet \
@ -47,6 +49,15 @@ RUN curl -sSL -O https://github.com/tesseract-ocr/tessdata_best/archive/$TESSDAT
rm -rf $TESSDATA_BEST_VERSION.tar.gz
# Build ocrd_olena
RUN curl -sSL -o ocrd_olena.tar.gz https://github.com/OCR-D/ocrd_olena/archive/fde4436.tar.gz && \
mkdir ocrd_olena && \
tar xvz -C ocrd_olena --strip-components=1 -f ocrd_olena.tar.gz && \
cd ocrd_olena && \
make install PREFIX=/usr/local && \
cd .. && rm -rf ocrd_olena ocrd_olena.tar.gz
# Copy OCR models
RUN mkdir -p /var/lib/calamari-models
COPY data/calamari-models/GT4HistOCR /var/lib/calamari-models/GT4HistOCR