DocCreator is an open source, cross-platform software allowing to generate synthetic document images and the accompanying groundtruth. Various degradation models can be applied on original document images to create virtually unlimited amounts of different images.
Citation: if you use DocCreator in Reaserch work for publication, please cite:
Journet, N.; Visani, M.; Mansencal, B.; Van-Cuong, K.; Billy, A.
DocCreator: A New Software for Creating Synthetic Ground-Truthed Document Images.
J. Imaging 2017, 3, 62.
http://www.mdpi.com/2313-433X/3/4/62
The program has the following dependencies :
- OpenCV (3.x or 4.x)
- Qt 5
- Tesseract [optional] If Tesseract is not found on the system, it will be compiled from (provided) sources. However, you will need to have network access during the configuration step to download tessdata, tesseract languages data.
- CMake is used for compilation configuration.
- Ninja [optional]. On Windows in particular, it may be convenient to install ninja to build all from the command line.
- C++ compiler This program should compile on linux (with gcc & clang), Mac OS (with clang) and Microsoft Windows 10 (with Visual Studio 2017). It has been tested on Fedora (19->30), Ubuntu (14.04->19.04), Mac OS (10.9->10.14), Windows 10.
On linux, your distribution may provide the binary packages for all required dependencies.
On Ubuntu (14.04 and above), you can install the required binary packages with the following command:
sudo apt-get install libopencv-dev qtbase5-dev qtdeclarative5-dev libqt5xmlpatterns5-dev cmake
You may also have to set Qt5 as default with:
sudo apt-get install qt5-default
You can install tesseract with the following command:
sudo apt-get install tesseract-ocr tesseract-ocr-fra libtesseract-dev libleptonica-dev
On Fedora (21 or 22), you can install the required binary packages with the following command as root:
yum install opencv-devel qt5-qtbase-devel qt5-qtxmlpatterns-devel cmake
On Fedora (23 and above), you can install the required binary packages with the following command as root:
dnf install opencv-devel qt5-qtbase-devel qt5-qtxmlpatterns-devel cmake gcc-c++
You can install tesseract with the following command as root:
dnf install tesseract tesseract-devel tesseract-langpack-fra leptonica leptonica-devel libpng-devel
To compile with clang on linux:
- you need to have libc++
On Ubuntu, you can install libc++ with the following command:
sudo apt-get install libc++-dev libc++abi-dev
Warning: on Ubuntu 16.04, you may have to edit /usr/include/c++/v1/string cf https://stackoverflow.com/questions/37096062/get-a-basic-c-program-to-compile-using-clang-on-ubuntu-16/38385967#38385967 On Fedora, you can use the following command:
dnf install libcxx - you need to have OpenCV 3.1 or above. See below how to build OpenCV from sources.
On linux, CMake is often provided by the distribution.
Binary packages of CMake last versions are available from https://cmake.org/download/
cmake must be in your PATH. You should be able to print cmake version with the following commande: cmake --version
[Optionnal]
Binary packages of Ninja are available from https://github.com/ninja-build/ninja/releases
ninja must be in your PATH. You should be able to print ninja version with the following commande: ninja --version
On OSX (10.9, 10.10 or 10.11) with Homebrew, you can install the required dependencies with the following commands:
brew install qt5
brew linkapps qt5
brew link --force qt5
brew tap homebrew/science
brew install opencv
On linux, it is recommanded to use packages provided by the distribution. See above for instructions.
On Mac and Windows, the open source version of Qt may be installed with the Qt Download Installer, downloaded from
https://www.qt.io/download
On Windows, you will have to choose the version to install according to the compiler that you use and the target architecture.
For example, "MSVC 2017 64-bit" if you use Microsoft Visual Studio 2017 on a 64-bit processor.
You can also install OpenCV from sources (if it is not provided by your distribution or if you want a newer version for example).
OpenCV is also built with CMake. You can use the following command to configure a minimal OpenCV version compatible with DocCreator:
cmake <PATH_TO_CMakeLists.txt> -DCMAKE_INSTALL_PREFIX=<INSTALL_PREFIX> -DBUILD_opencv_highgui=ON -DENABLE_FAST_MATH=ON -DBUILD_DOCS=OFF -DBUILD_PERF_TESTS=OFF -DBUILD_TESTS=OFF -DWITH_CUDA=OFF -DWITH_CUFFT=OFF -DWITH_FFMPEG=OFF -DWITH_GIGEAPI=OFF -DWITH_JASPER=OFF -DWITH_LIBV4L=OFF -DWITH_MATLAB=OFF -DWITH_OPENCL=OFF -DWITH_OPENCLAMDBLAS=OFF -DWITH_OPENCLAMDFFT=OFF -DWITH_OPENEXR=OFF -DWITH_PVAPI=OFF -DWITH_V4L=OFF -DWITH_VTK=OFF -DWITH_WEBP=OFF -DWITH_1394=OFF -DBUILD_opencv_apps=OFF -DBUILD_opencv_calib3d=OFF -DBUILD_opencv_dnn=OFF -DBUILD_opencv_features2d=OFF -DBUILD_opencv_flann=OFF -DBUILD_opencv_ml=OFF -DBUILD_opencv_objdetect=OFF -DBUILD_opencv_shape=OFF -DBUILD_opencv_stitching=OFF -DBUILD_opencv_superres=OFF -DBUILD_opencv_ts=OFF -DBUILD_opencv_video=OFF -DBUILD_opencv_videoio=ON -DBUILD_opencv_videostab=OFF -DBUILD_opencv_world=OFF -DBUILD_opencv_highgui=ON -DPYTHON2_EXECUTABLE="" -DPYTHON3_EXECUTABLE="" -DBUILD_opencv_python2=OFF -DBUILD_opencv_python3=OFF -DWITH_QUICKTIME=OFF -DHAVE_QTKIT=FALSE
with corrected <PATH_TO_CMakeLists.txt> and <INSTALL_PREFIX>.
CMake can generate a .sln solution file to open with Visual Studio.
First, open a "Developer Command Prompt" ("Developer Command Prompt for VS 2017" for VS17 for example).
cd <OpenCV_SOURCES_DIRECTORY>
mkdir build
cd build
For Visual Studio 2017 x64, you can use the following command to generate a .sln file:
cmake <PATH_TO_CMakeLists.txt> -G "Visual Studio 15 2017 Win64" -DCMAKE_INSTALL_PREFIX=<INSTALL_PREFIX> -DBUILD_opencv_highgui=ON -DENABLE_FAST_MATH=ON -DBUILD_DOCS=OFF -DBUILD_PERF_TESTS=OFF -DBUILD_TESTS=OFF -DWITH_CUDA=OFF -DWITH_CUFFT=OFF -DWITH_FFMPEG=OFF -DWITH_GIGEAPI=OFF -DWITH_JASPER=OFF -DWITH_LIBV4L=OFF -DWITH_MATLAB=OFF -DWITH_OPENCL=OFF -DWITH_OPENCLAMDBLAS=OFF -DWITH_OPENCLAMDFFT=OFF -DWITH_OPENEXR=OFF -DWITH_PVAPI=OFF -DWITH_V4L=OFF -DWITH_VTK=OFF -DWITH_WEBP=OFF -DWITH_1394=OFF -DBUILD_opencv_apps=OFF -DBUILD_opencv_calib3d=OFF -DBUILD_opencv_dnn=OFF -DBUILD_opencv_features2d=OFF -DBUILD_opencv_flann=OFF -DBUILD_opencv_ml=OFF -DBUILD_opencv_objdetect=OFF -DBUILD_opencv_shape=OFF -DBUILD_opencv_stitching=OFF -DBUILD_opencv_superres=OFF -DBUILD_opencv_ts=OFF -DBUILD_opencv_video=OFF -DBUILD_opencv_videoio=ON -DBUILD_opencv_videostab=OFF -DBUILD_opencv_world=OFF -DBUILD_opencv_highgui=ON -DPYTHON2_EXECUTABLE="" -DPYTHON3_EXECUTABLE="" -DBUILD_opencv_python2=OFF -DBUILD_opencv_python3=OFF -DWITH_QUICKTIME=OFF -DHAVE_QTKIT=FALSE
Then open the .sln file with Visual Studio,
check that you have "Release" and "x64" for the "Solution configurations" and "Solution platforms" (comboboxes in the toolbar),
then go to the menu "Build" and do "Build solution".
To install OpenCV in the target directory (<INSTALL_PREFIX>), in the "Solution explorer" (on the right), right click on "INSTALL" in CMakeTargets and choose "Build".
It will install OpenCV in <INSTALL_PREFIX>, and in particular .lib files and OpenCVConfig.cmake should be in <INSTALL_PREFIX>/x64/vc15/lib and .dll files should be in <INSTALL_PREFIX>/x64/vc15/bin.
CMake can generate a ninja build. The whole compilation is done from the command line.
First, open a "Native Tools Command Prompt" ("x64 Native Tools Command Prompt for VS 2017" for VS17 on x86-64 for example).
cd <OpenCV_SOURCES_DIRECTORY>
mkdir build
cd build
cmake <PATH_TO_CMakeLists.txt> -G "Ninja" -DCMAKE_INSTALL_PREFIX=<INSTALL_PREFIX> -DBUILD_opencv_highgui=ON -DENABLE_FAST_MATH=ON -DBUILD_DOCS=OFF -DBUILD_PERF_TESTS=OFF -DBUILD_TESTS=OFF -DWITH_CUDA=OFF -DWITH_CUFFT=OFF -DWITH_FFMPEG=OFF -DWITH_GIGEAPI=OFF -DWITH_JASPER=OFF -DWITH_LIBV4L=OFF -DWITH_MATLAB=OFF -DWITH_OPENCL=OFF -DWITH_OPENCLAMDBLAS=OFF -DWITH_OPENCLAMDFFT=OFF -DWITH_OPENEXR=OFF -DWITH_PVAPI=OFF -DWITH_V4L=OFF -DWITH_VTK=OFF -DWITH_WEBP=OFF -DWITH_1394=OFF -DBUILD_opencv_apps=OFF -DBUILD_opencv_calib3d=OFF -DBUILD_opencv_dnn=OFF -DBUILD_opencv_features2d=OFF -DBUILD_opencv_flann=OFF -DBUILD_opencv_ml=OFF -DBUILD_opencv_objdetect=OFF -DBUILD_opencv_shape=OFF -DBUILD_opencv_stitching=OFF -DBUILD_opencv_superres=OFF -DBUILD_opencv_ts=OFF -DBUILD_opencv_video=OFF -DBUILD_opencv_videoio=ON -DBUILD_opencv_videostab=OFF -DBUILD_opencv_world=OFF -DBUILD_opencv_highgui=ON -DPYTHON2_EXECUTABLE="" -DPYTHON3_EXECUTABLE="" -DBUILD_opencv_python2=OFF -DBUILD_opencv_python3=OFF -DWITH_QUICKTIME=OFF -DHAVE_QTKIT=FALSE
then build with:
ninja
and install with:
ninja install
You then have to set the OPENCV_DIR environment variable for OpenCV to be found by CMake.
You can open a cmd prompt "CMD.exe" with Administrator privileges (right click), and do:
setx -m OPENCV_DIR <INSTALL_PREFIX>
CMake is used for configuration & compilation. The following commands allow to build DocCreator.
On Linux and Mac, type the following commands from a terminal:
cd <DocCreatorDirectory>
mkdir build
cd build
cmake .. -DCMAKE_INSTALL_PREFIX=<MyInstallationPrefix>
make
make install
You can also pass other options to cmake. See below.
You can then launch the executable:
<MyInstallationPrefix>/bin/DocCreator
On Windows, with Visual Studio:
First, open a "Developer Command Prompt" ("Developer Command Prompt for VS 2017" for VS17 for example).
(you can check that OPENCV_DIR is correctly set by doing: echo %OPENCV_DIR%)
go to the source directory,
cd <DocCreatorDirectory>
mkdir build
cd build
CMake can generate a .sln solution file to open with Visual Studio.
For Visual Studio 2017:
cmake .. -G "Visual Studio 15 2017 Win64" -DCMAKE_PREFIX_PATH=<QT_PATH> -DCMAKE_INSTALL_PREFIX=<MyInstallationPrefix>
or for Visual Studio 2019 (with CMake 3.14.6 or above, in "Developer Command Prompt for VS 2019"):
cmake .. -G "Visual Studio 16 2019" -A "x64" -DCMAKE_PREFIX_PATH=<QT_PATH> -DCMAKE_INSTALL_PREFIX=<MyInstallationPrefix>
<QT_PATH> is the path to installed Qt configuration for this compiler. <QT_PATH> is for example C:/Qt/5.12.4/msvc2017_64 You can also pass other options to cmake. See below.
Then you can open the generated .sln file with Visual Studio. Check that you have "Release" and "x64" for the "Solution configurations" and "Solution platforms" (comboboxes in the toolbar), then go to the menu "Build" and do "Build solution". To install DocCreator in the target directory (), in the "Solution explorer" (on the right), right click on "INSTALL" in CMakeTargets and choose "Build".
CMake can generate a ninja build. The whole compilation is done from the command line.
cmake .. -G "Ninja" -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH=<QT_PATH> -DCMAKE_INSTALL_PREFIX=<MyInstallationPrefix>
You can also pass other options to cmake. See below.
then build with:
ninja
then install with:
ninja install
You can then launch the executable:
<MyInstallationPrefix>/programs/DocCreator.exe
When configuring DocCreator with cmake, you can pass the option -DBUILD_OTHER_PROGS=ON to cmake. It will be build other example programs using DocCreator framework.
In particular, it will build:
* DocCreatorDegradator that allows to apply degradation effects on all the images of a given directory and save produced images in a new directory. You can change the applied degradation effects in software/DocCreator/src/Degradator/main.cpp
* DocCreatorMakeFont that allows to create DocCreator old-fonts from existing (TrueType) fonts. You can save the produced files in data/font to use these fonts in DocCreator.
When configuring DocCreator with cmake, you can pass the option -DBUILD_OTHER_PROGS_3D=ON to cmake. It will be build other example programs using DocCreator 3D code.
When configuring DocCreator with cmake, you can pass the option -DBUILD_TESTING=ON to cmake. It will be build degradations unit tests.
See FAQ.md.
-
If you encounter a problem with Qt during the cmake process, ensure that the qmake executable is in your PATH You maye have to do:
export PATH=/Users/XXX/Qt/5.6/clang_64/bin:$PATH
or
export PATH=/usr/local/opt/qt5/bin:$PATH
before executingcmake .. -
If you encounter a problem with OpenCV during the cmake process, you may have to specify the directory containing the file OpenCVConfig.cmake For example, on linux and macOS:
OpenCV_DIR=/Users/XXX/tools/share/OpenCV cmake ..
On Windows, to set the OpenCV_DIR environment variable, you may open a cmd prompt "CMD.exe" with Administrator privileges (right click), and do:
setx -m OPENCV_DIR <INSTALL_PREFIX>
You will have to open a new terminal. You can print and check that the variable is correctly set with:
echo %OpenCV_DIR% -
On Windows, if the compiler is not correctly found by cmake
you may have more informations in the file build/CMakeFiles/CMakeError.log
In particular, if ucrtd.lib is not found, you may have to install "Windows SDK".
Last version of Windows 10 SDK should be available from https://developer.microsoft.com/en-US/windows/downloads/windows-10-sdk (tested with version 10.0.15063.468). -
If CMake prints the following error:
"Found OpenCV Windows Pack but it has no binaries compatible with your configuration."
it may indicate that you have a mismatch between the architectures of OpenCV libs and your compiler.
For example, OpenCV libs may have been build for 64-bit architecture but your are trying to compile for 32-bit. On Windows with Visual Studio in particular, if you use a "Native Tools Command Prompt", check that it is for the correct architecture. For example "x64 Native Command Prompt" and not "x86 Native Command Prompt" for x86_64. -
If CMake fails to download one of the dependencies and prints this kind of error:
error: downloading
https://gitlab.com/libeigen/eigen/-/archive/3.3.7/eigen-3.3.7.tar.gz
failed: status code=1 string="Unsupported protocol", log=Protocol "https"
not supported or disabled in libcurl
the problem seems to be that the CURL library shipped with CMake is not build with SSL support. You may have to compile CMake from sources enabling using a CURL library with SSL support.
-
If you encounter a problem during the make process, it may be related to the C++11 detection edit CMakeLists.txt and add (or remove) the following line before the first "AD_SUBDIRECTORY" line
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
then restart the construction process:
rm -rf CMakeCache.txt CMakeFiles
cmake ..
make -
On Windows, with Visual Studio, if you encounter a link problem with tesseract, it may be that you compiled in Debug mode instead of Release mode. Indeed, DocCreator tries to link with tesseract305.lib, but tesseract305d.lib is generated in Debug mode.
-
On Windows, if you encounter a problem when the program is launched, it may be that the various required libraries (dll) are not correctly found.
You then may have to copy some files to /programs directory [if they are not present]:
- from OpenCV <INSTALL_PREFIX>/x64/vc15/bin :
opencv_coreXXX.dll, opencv_highguiXXX.dll, opencv_imgcodecsXXX.dll, opencv_imgprocXXX.dll, opencv_photoXXX.dll, opencv_videoioXXX.dll (XXX has to be replaced by 330 for OpenCV 3.3) - from Qt:
- from <QT_PATH>/bin :
Qt5Core.dll, Qt5Gui.dll, Qt5Network.dll, Qt5OpenGL.dll, qt5PrintSupport.dll, Qt5Widgets.dll, Qt5Xml.dll, Qt5XmlPatterns.dll, libEGL.dll, libGLESv2.dll, d3dcompiler_47.dll - from <QT_PATH>/plugins :
imageformats directory platforms directory
- from <QT_PATH>/bin :