DIPlib is a library for quantitative image analysis. It has been in development at Delft University of Technology in The Netherlands since 1995. The 3.0 release represents a complete rewrite in modern C++ of the library infrastructure, with most of the image processing and analysis algorithms ported unmodified from the previous version.
See the documentation for more information.
Building the library
Linux, MacOS, Cygwin and other Unix-like systems
To build the library you will need a C++14 compliant compiler and CMake. See below under "Dependencies" for optional dependencies that you can install to improve your DIPlib experience. Use the following commands to build:
mkdir target cd target cmake /path/to/dip/root/directory make -j install
-j option to
make enables a multi-threaded build. Limit the number of
concurrent jobs to, for example, 4 with
For step-by-step instructions for MacOS, see
all (default) # builds DIPlib, DIPimage and PyDIP, if configured install # builds and installs target 'all' check # builds the unit_tests program and runs it check_memory # ...and runs it under valgrind apidoc # builds the HTML documentation for the library API examples # builds the examples package # creates a distributable package
make targets are part of the
DIP # builds the DIPlib library DIPviewer # builds the DIPviewer module (plus the DIPlib library) PyDIP # builds the PyDIP Python module (includes DIP and DIPviewer targets) dum # builds the DIPimage User Manual PDF
cmake command-line arguments:
-DCMAKE_INSTALL_PREFIX=$HOME/dip # choose an instal location -DCMAKE_BUILD_TYPE=Debug # by default it is release -DDIP_SHARED_LIBRARY=Off # build a static DIPlib library -DCMAKE_C_COMPILER=gcc-6 # specify a C compiler (for libics) -DCMAKE_CXX_COMPILER=g++-6 # specify a C++ compiler (for everything else) -DCMAKE_CXX_FLAGS="-march=native" # specify additional C++ compiler flags -DDIP_ENABLE_STACK_TRACE=Off # disable stack trace generation on exception -DDIP_ENABLE_ASSERT=Off # disable asserts -DDIP_ENABLE_DOCTEST=Off # disable doctest within DIPlib -DDIP_ENABLE_MULTITHREADING=Off # disable OpenMP multithreading -DDIP_ENABLE_ICS=Off # disable ICS file format support -DDIP_ENABLE_TIFF=Off # disable TIFF file format support -DDIP_ENABLE_FFTW=On # enable the use of FFTW3 -DDIP_ENABLE_UNICODE=Off # disable UFT-8 strings within DIPlib -DDIP_ALWAYS_128_PRNG=On # use the 128-bit PRNG code where 128-bit # integers are not natively supported -DDIP_BUILD_DIPVIEWER=Off # don't build/install the DIPviewer module -DDIP_BUILD_PYDIP=Off # don't build/install the PyDIP Python module -DDIP_BUILD_DIPIMAGE=Off # don't build/install the DIPimage MATLAB toolbox -DPYBIND11_PYTHON_VERSION=3.6 # compile PyDIP agains Python 3.6 -DMatlab_ROOT_DIR=<path> # compile DIPimage against MATLAB in <path>
Some of these options might not be available on your system. For example, if you don't have
MATLAB installed, the
DIP_BUILD_DIPIMAGE option will not be defined. In this case, setting
Off will yield a warning message when running CMake.
Note that on some platforms, the Python module requires the DIPlib library to build as
a dynamic load library (
-DDIP_SHARED_LIBRARY=On, which is the default).
apidoc target requires that Doxygen be installed, the target will not be available
if it is not. The
dum target requires that Pandoc be installed, the target will not be
available if it is not; this target will fail to build if additional tools are not installed
(see below under "Dependencies").
Unless you want to use Cygwin or MinGW (see above), we recommend Microsoft Visual Studio 2017 (version 15). You'll also need CMake.
Using CMake-gui, choose where the source directory is and where to build the binaries. Then press "Configure" and select Visual Studio. Finally, press "Generate". You should now have a Visual Studio solution file that you can open in Visual Studio and build as usual.
For step-by-step instructions, see
See below for optional dependencies that you can install to improve your DIPlib experience.
Below we list all external dependencies needed to compile the various parts of the project. DIPlib also depends on a few other external projects, whose sources are included in this repository (see below under "License" for more information). Note that, with the exception of dynamic linking to a few external libraries, none of these dependencies are required when using the DIPlib library (that is, DIPlib's public header files do not import headers from other projects).
If you have FFTW3 installed, you can set the
CMake variable to have DIPlib use FFTW3 instead of the built-in FFT algorithm.
FFTW3 is more efficient, especially for image sizes that do not factor into small
numbers, but it has a copyleft license.
DIPimage requires that MATLAB be installed
for compilation and execution (of course).
Optionally, you can install OME Bio-Formats to
enable DIPimage to read many microscopy image file formats (type
help readim in MATLAB,
after installing DIPimage, to learn more).
PyDIP requires that Python (preferably Python3) be installed.
To build the DIPlib documentation, Doxygen is needed. There is a chance it will only work on Unix-like systems (not yet tested under Windows).
The DIPimage User Manual requires Pandoc,
LaTeX. Note that you'll need certain LaTeX packages,
upquote, that are not in the most basic set of packages. You can install these
through the TeX Live package manager.
Linking against the library
When using CMake, and importing the
DIP target into your project in the right way, you will just need
to link against the
DIP target and everything will be configured correctly. Otherise, there are several
macros that you should define when building any program that links against DIPlib:
If DIPlib was build with the
DIP_SHARED_LIBRARY flag was not set, then you need to define the
macro when compiling the code that links against it. Likewise, if the
DIP_ALWAYS_128_PRNG flag was set,
then you must define a
DIP__ALWAYS_128_PRNG macro when compiling your program. Mismatching this flag
could cause your program to not link, or worse, crash at runtime.
The following flags do not need to be matched, but they should be if you want the inline functions to behave the same as the pre-compiled ones:
For DIPviewer, if
DIP_SHARED_LIBRARY was not set, define the
DIP__VIEWER_IS_STATIC macro. Also define
DIP__HAS_GLFW depending on which back-end is used.
Copyright 2014-2018 Cris Luengo
Copyright 1995-2014 Delft University of Technology
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this library except in compliance with the License. You may obtain a copy of the License at
(or see the LICENSE.txt file in this distribution)
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
DIPlib includes the following libraries / external code:
Mozilla Public License Version 2.0
(we do not use any of the components with more restrictive licenses)
DocTest (its use can be disabled)
Copyright 2016-2017 Viktor Kirilov
The MIT License (MIT)
The DFT algorithm out of OpenCV 3.1
Copyright 2000 Intel Corporation
Intel License Agreement For Open Source Computer Vision Library
PCG Random Number Generation for C++
Copyright 2014-2017 Melissa O'Neill and the PCG Project contributors
Apache 2.0 License, or The MIT License, at your option
A fast 2D labeling algorithm out of YACCLAB
Copyright 2016-2017 Costantino Grana, Federico Bolelli, Lorenzo Baraldi and Roberto Vezzani
3-Clause BSD License
A few color maps from colorcet Copyright 2017 Peter Kovesi 1-Clause BSD-like License see src/display/colormap.cpp
libics (its use can be disabled)
Copyright 2015-2017 Scientific Volume Imaging Holding B.V.
Copyright 2000-2013 Cris Luengo and others
GNU Lesser General Public License, Version 2.1
pybind11 (only used in the Python bindings)
Copyright 2016 Wenzel Jakob
3-Clause BSD License
A few MATLAB scripts from OME Bio-Formats
Copyright 2012-2017 Open Microscopy Environment
GNU General Public License, Version 2
(the full Bio-Formats library needs to be installed for these to be useful)
Note that all of these have permissive open-source licenses similar in spirit to the Apache License, except for OME Bio-Formats.
DIPlib also optionally links against:
LibTIFF (as installed on your system)
Copyright 1988-1997 Sam Leffler
Copyright 1991-1997 Silicon Graphics, Inc.
ZLib (as installed on your system)
Copyright 1995-2017 Jean-loup Gailly and Mark Adler
FFTW3 (as installed on your system, not used by default)
Copyright (c) 2003, 2007-14 Matteo Frigo
Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
GNU General Public License, Version 2 (not compatible with the Apache License)
DIPviewer links against the following libraries:
FreeGLUT (as installed on your system, alternative to GLFW)
Copyright 1999-2000 Pawel W. Olszta
GLFW (as installed on your system, alternative to FreeGLUT)
Copyright 2002-2006 Marcus Geelnard
Copyright 2006-2011 Camilla Berglund
OpenGL (as installed on your system)
(free from licensing requirements)
GLEW (as installed on your system, for Windows only)
Copyright 2008-2016, Nigel Stewart
Copyright 2002-2008, Milan Ikits
Copyright 2002-2008, Marcelo E. Magallon
Copyright 2002, Lev Povalahev
Modified BSD License