MorphoGraphX runs on both Linux and Windows. However, we develop and use it mostly on Linux, so the Linux version better tested.
Some MorphoGraphX versions use the Cuda toolkit, which requires a 3D nVidia graphics card with enough memory to handle the size of stacks being processed. For large datasets we recommend a PC with 32-64Gb of Ram and a decent nVidia graphics card. If you have troubles running the Cuda version of MorphoGraphX, be sure that your nVidia graphics driver version is up to date and new enough to support your graphics card.
MorphoGraphX 2.0
MorphoGraphX 2.0 adds the ability to annotate samples with positional information, and many new tools for 2.5D and 3D image processing, including machine learning tools for voxel and cell classification. The Cuda versions of the packages below were compiled with Cuda 11.4.
Linux
(Ubuntu 20.04)
The following package can be installed on Ubuntu 20.04 and was compiled with Cuda 11.4. On Ubuntu 20, HDF5 files are supported: MGX-2.0.1-254-Ubuntu20.04-Cuda11
The following package can be installed on Ubuntu 20.04. It does not require Cuda or an nVidia graphics card: MGX-2.0.1-249-Ubuntu20.04-OMP
This package can be installed after you have installed the Cuda version of MGX above, and it includes processes for stack processing by convolutional neural networks to improve segmentation as described here: https://arxiv.org/pdf/1810.06933.pdf. It requires Cuda and you might need to additionally install the libcudnn runtime available from the nVidia website. This package is quite large, as it is a “fat” binary compiled for multiple Cuda architectures: MGX-2.0.1-254-Ubuntu20.04-Cuda11-CNN
To use ITK processes, you will also need the ITK library. The packages above were compiled with ITK 5.1.2 with the use of the system HDF5 enabled. You can build your own, or download and install the following package: ITK-5.1.2-Linux
(Ubuntu 18.04)
The following package can be installed on Ubuntu 18.04 and was compiled with Cuda 11.4: MGX-2.0.1-256-Ubuntu18.04-Cuda11
The following package can be installed on Ubuntu 18.04. It does not require Cuda or an nVidia graphics card: MGX-2.0.1-248-Ubuntu18.04-OMP
This package can be installed after you have installed the Cuda version of MGX above, and it includes processes for stack processing by convolutional neural networks to improve segmentation as described here: https://arxiv.org/pdf/1810.06933.pdf. It requires Cuda and you might need to additionally install the libcudnn runtime available from the nVidia website. This package is quite large, as it is a “fat” binary compiled for multiple Cuda architectures: MGX-2.0.1-256-Ubuntu18.04-Cuda11-CNN
To use ITK processes, you will also need the ITK library. The packages above were compiled with ITK 5.0.1. You can build your own, or download and install the following package: ITK-5.0.1-Linux
Windows
This package can be installed on Windows 10. It does not require Cuda or an nVidia graphics card: MGX-2.0.1-246-Windows
Source Code
If you have a different version of Linux or Cuda, you can compile MorphoGraphX from source: mgxsource-2.0.1-259
MorphoGraphX Plugins and Scripts
MorphoGraphX can be extended with plugins written in C++ and loaded at runtime. It is also possible to run python scripts to do batch processing from within MorphoGraphX.
Plugin Examples
Sample plugins (to help you get started writing your own) are here:
MGXSamplePlugins.zip
Python Script
The following Python script can be used to combine lineages from multiple time points:
LineageCombine
MorphoGraphX 1.0
This section contains Linux and Windows packages for the older versions of MorphoGraphX. All of these packages require Cuda and an nVidia graphics card. It is also possible to compile an older version from the source code.
Linux
The following package can be installed on Linux Mint 19 or Ubuntu 18.04 with Cuda 9.1:
MGX-1.0.1280-LinuxMint19-CellAtlas-Cuda9.1.deb
The following package can be installed on Linux Mint 18 or Ubuntu 16.04 with Cuda 9.1:
MGX-1.0.1280-LinuxMint18.1-CellAtlas-Cuda9.1.deb
The following package can be installed on Linux Mint 17 or Ubuntu 14.04 with Cuda 5:
MGX-LinuxMint17-1.0.1280-CellAtlas.deb
The following package can be installed on Linux Mint 17 or Ubuntu 14.04 with Cuda 6.5:
MGX-LinuxMint17-1.0.1280-CellAtlas-Cuda6.5.deb
To use ITK processes, you will also need the ITK library. The packages above were compiled with ITK 4.5.0. You can build your own, or download and install the following package: ITK-4.5.0-Linux.deb
Windows
Please download this Windows installer for Window 7 or 8 (64bit):
MorphoGraphX-1.1.1280-CellAtlas.exe
Before installing MorphoGraphX, please go to the nVidia website and install Cuda 7.0. You can test your Cuda installation by running one of their example programs.
MorphoGraphX 