This is a Blender add-on to generate a mesh from 3 images describing its silhouette from each axis.

Load the images you want to generate a mesh from into your project. Clear your selection and add a new Shadowbox from Add > Shadowbox. In the settings panel in the lower left corner, assign each axis an image. The images will we stretched to fit a unit cube, but you can scale the box to the wanted aspect ratio.

Every time you execute the Shadowbox function on a mesh object, that object's geometry will be overwritten. If you don't want to loose any of your meshes, execute the function without a selection and a new object will be created.

Because high mesh resolutions will quickly drain performance, it is recommended to only work with the resolution you really need. When you decided on one you can slightly improve performance by manually scaling your images to the wished resolution. Then the add-on doesn't have to create a temporary copy of your images and resize it every time you regenerate the mesh.

This is especially useful when you use the Run in background option, which regenerates the mesh every frame while you can paint into the images from inside Blender's Image Editor. Stop the automatic update by pressing Escape.

Download the archive from the Releases page. Inside Blender, navigate to Edit > Preferences > Add-ons and click the Install button. Choose the downloaded archive and confirm.

Under Linux you also have to install tbb as thread scheduler. For example under Ubuntu: sudo apt-get install libtbb12.

For optimal performance this add-on uses a custom Python module written in C which has to be compiled for the Python version the targeted Blender version ships with. CMake is used to unify the build process regardless of your operating system, but the way you install dependencies varies. The module has the following dependencies:

• openVDB
• pybind11
• eigen3

Follow the instructions under Windows or Linux depending on your operating system. The end product is a file with the extension .pyd under Windows and .so under Linux.

• Copy it into the shadowbox directory at the root of this repository
• Copy that directory into Blender's add-on directory
• In Blender, navigate to Edit > Preferences > Add-ons
• Hit the Refresh button
• Search for the add-on and tick it in the list

For Blender version 3.4:

• Windows: %APPDATA%/Blender Foundation/Blender/3.4/scripts/addons
• Ubuntu: ~/.config/blender/3.4/scripts/addons

If still unsure:

• Press Shift F4 inside Blender to open the Python console
• Execute bpy.utils.user_resource('SCRIPTS', path="addons") to see the path

It is recommended to install dependencies with vcpkg. Assuming it has been installed to C:/vcpkg, execute the following commands where you want to build Shadowbox:

vcpkg install eigen3 openvdb --triplet=x64-windows-static
vcpkg install pybind11 --triplet=x64-windows
git clone https://github.com/D4KU/shadowbox.git
cd shadowbox
mkdir build
cd build
cmake .. -DCMAKE_TOOLCHAIN_FILE=C:/vcpkg/scripts/buildsystems/vcpkg.cmake -DVCPKG_TARGET_TRIPLET=x64-windows-static -Dpybind11_DIR=C:/vcpkg/installed/x64-windows/share/pybind11
cmake --build . --config Release

Because the static OpenVDB build from vcpkg seems to be broken, I had to compile OpenVDB from source as well. This may be resolved in the future. In case that happens to you, here are my build instructions. The last command for the installation you have to execute on a command line with admin rights.

vcpkg install boost tbb --triplet=x64-windows-static
git clone https://github.com/AcademySoftwareFoundation/openvdb
cd openvdb
mkdir build
cd build
cmake .. -DCMAKE_TOOLCHAIN_FILE=C:/vcpkg/scripts/buildsystems/vcpkg.cmake -DVCPKG_TARGET_TRIPLET=x64-windows-static -DUSE_BLOSC=OFF -DUSE_ZLIB=OFF -DOPENVDB_CORE_SHARED=OFF -A x64
cmake --build . --config Release --target install

Before proceeding to build Shadowbox I had to rename libopenvdb.lib in the installation directory to openvdb.lib.

sudo apt-get install libeigen3-dev pybind11-dev libopenvdb-dev
git clone https://github.com/D4KU/shadowbox.git
cd shadowbox
mkdir build
cd build
cmake ..
cmake --build . --config Release

sudo apt-get install libboost-iostreams-dev libtbb-dev
git clone https://github.com/AcademySoftwareFoundation/openvdb
cd openvdb
mkdir build
cd build
cmake .. -DUSE_BLOSC=OFF -DUSE_ZLIB=OFF -DOPENVDB_CORE_SHARED=OFF -DCMAKE_POSITION_INDEPENDENT_CODE=ON
sudo cmake --build . --config Release --target install