Building » Building the DIPlib project on Windows

Contents

Compiling DIPlib requires a few programs that do not come preinstalled on Windows. We describe here how to obtain these programs, and how to use them to compile the DIPlib project.

See CMake configuration for additional information on the build targets and CMake configuration options.

Visual Studio

You can download the free MS Visual Studio Community Edition from its website.

We recommend that you use at least the 2019 edition. We were able to build with the 2015 and 2017 editions, but had to turn off some components that would not build because they used some advanced C++14 constructs that will not compile with older versions of MSVC.

Download and install as directed. Select the “Desktop C++ applications” option. Optionally, you can select to install Python 3 as well. You can do this if you don’t have Python but want to compile the PyDIP interface.

Make sure you select the Git option as well. If you don’t have Git installed, and cannot install it through Visual Studio, download and install it from here.

CMake

You can obtain the latest CMake on its website.

Again, download and install as directed.

dox++

To compile the documentation yourself (which shouldn’t be necessary, but some IDEs might require), you need dox++. See Building the DIPlib documentation for details.

Cloning the repository

Next, get the source repository from GitHub. Open the “Git GUI” program on your start menu, and select “Clone existing repository”. The source location is

https://github.com/DIPlib/diplib.git

Pick any directory on your system as destination. For example src\DIPlib in your user directory. The standard clone type is OK to use. Click “Clone”.

Installing dependencies

Windows does not have a simple dependency installation system, so this step involves a lot of manual labor. Fortunately, all these dependencies are optional, so feel free to skip this section.

  • Download the GLFW binaries. If you intend to build a 64-bit version of DIPlib (recommended!) get the 64-bit version of GLFW. Extract the ZIP file and note the location.

  • Download the Bio-Formats library. Put it somewhere sensible and note the location. The same location (as well as the library installation path) will be used to find it during execution.

    If building only DIPimage (the MATLAB toolbox), you don’t need to download Bio-Formats before building; instead follow the directions you can read when you do help readim in MATLAB after installation.

Creating Visual Studio project files

Open the CMake program. Enter the name of the directory you cloned the repository in (in our example, C:\Users\<name>\src\DIPlib); you can browse to this directory. Under “Where to build the binaries” enter a different, new directory. For example target\DIPlib in your user directory. Click on “Configure”. A pop-up window will ask you for which generator to use. You should select your version of Visual Studio here. Make sure you select the Win64 version. If you select the default version, you will build 32-bit binaries (really, in 2019 we’re still building 32-bit binaries by default?). Also, the latest MATLAB versions no longer support 32-bit binaries, so unless you select the 64-bit generator, you won’t be able to build DIPimage.

You will see a series of options, and a configuration report at the bottom. Here you can change options:

  • CMAKE_INSTALL_PREFIX should be set to some directory where you want DIPlib and friends installed. By default this is C:\Program Files\DIPlib. You can pick any directory to your liking here. Note that the default directory requires administrator privileges (I was not able to install there, even though I have the administrator password). Let’s say we select C:\Users\<name>\DIPlib.

  • If DIP_BUILD_DIPIMAGE is not on the list, you don’t have MATLAB installed, or you are building 32-bit binaries.

  • If DIP_BUILD_DIPVIEWER is not on the list, you need to specify the locations for GLFW. Click the “Advanced” check box, this will show additional parameters. Look for GLFW_INCLUDE_DIR and GLFW_LIBRARY. For both of these, click the “…” button at the right and navigate to where you extracted the GLFW ZIP file. Select the “include” directory for the first parameter, and “lib-vc2015\glfw3.lib” for the second.

  • If DIP_BUILD_DIPVIEWER_JAVA is not on the list but DIP_BUILD_DIPVIEWER is, you don’t have MATLAB installed, you are building 32-bit binaries, or the Java SDK could not be found. This is only necessary if you want to use the viewslice command from DIPimage.

  • If DIP_BUILD_JAVAIO is not on the list, the Java SDK could not be found. This is only necessary if you want to import image formats that are not directly supported by DIPlib, and only for Python and C++ (MATLAB uses the Bio-Formats package differently, see help readim after installation).

  • If using a version of MSVC older than the 2019 edition, set both DIP_ENABLE_UNICODE and DIP_ENABLE_DOCTEST to Off. You will get compilation errors if you don’t do this.

  • If you set DIP_SHARED_LIBRARY to Off, DIPimage and PyDIP will likely not work correctly.

Finally, click on “Generate” to create a Visual Studio solution file.

Building

Find the file DIPlib.sln in the directory you selected as output for CMake, and open it. Also, CMake will prompt you to open this file after it has generated it.

Opening the solution file will launch Visual Studio. In the “Solution Explorer” you’ll find a list of targets (if configured for DIPimage, there will be very many targets!). The DIP target is the DIPlib library itself. To build the DIPimage toolbox, use the INSTALL target. This target builds everything and installs it in the destination directory you specified in CMake (C:\Users\<name>\DIPlib in our example). This is the target you will want to build. To additionally install PyDIP, use the pip_install target (but note that you need to run the INSTALL target first).

In the tool bar, make sure that “Release” and “x64” are selected (or x86 if you want to build 32-bit binaries). Right-click on INSTALL and select “Build”.

If everything works correctly, you will have (depending on which options were enabled during configuration):

  • C:\Users\<name>\DIPlib\bin: DIP.dll, as well as DIPviewer.dll, DIPjavaio.dll, DIPjavaio.jar, dipview.exe and dipviewjava.exe.

  • C:\Users\<name>\DIPlib\lib: DIP.lib, as well as DIPviewer.lib and DIPjavaio.lib.

  • C:\Users\<name>\DIPlib\include: The DIPlib include files, which you’ll need when building your own C++ programs using DIPlib.

  • C:\Users\<name>\DIPlib\share\DIPimage: The DIPimage toolbox for MATLAB.

  • In your selected Python package path: diplib (the Python module).

Using DIPimage

Once the INSTALL target has finished building and installing the toolbox, start MATLAB. Type the following commands:

addpath('C:\Users\<name>\DIPlib\share\DIPimage')
setenv('PATH',['C:\Users\<name>\DIPlib\bin',';',getenv('PATH')]);

This will make the toolbox available (replace C:\Users\<name>\DIPlib with the actual path you installed to).

To get started using DIPimage, read the DIPimage User Manual, and look through the help, starting at

help DIPimage

Or start the GUI:

dipimage

Using PyDIP

Once the pip_install target has finished installing, start Python. The following command will import the PyDIP package as dip, which is shorter to type and mimics the namespace used in the C++ library:

import diplib as dip

To get started using PyDIP, look through the help, starting at

help(dip)

The PyDIP User Manual is still quite short, but does contain some important information to get you started.

« Building the DIPlib project on macOS | Building | Building the DIPlib documentation »