pyvXRAY

An ABAQUS plug-in for the creation of virtual x-rays from 3D finite element bone/implant models.

Developed together with bonemapy and BMDanalyse to provide tools for preparation and post-processing of bone/implant computer models.

Copyright 2013, Michael Hogg ([email protected])

MIT license - See LICENSE.txt for details on usage and redistribution

Requirements

Software requirements

• ABAQUS >= 6.11 (tested on 6.11, 6.12 and 6.13)
• Python Image Library (PIL) >= 1.1.6 OR Pillow >= 2.2.0 (Note: Pillow is in active development and is recommended over PIL)

For building cython modules from source (e.g. if not using releases with pre-built modules):

• A C compiler. Using ABAQUS Python on Windows requires Microsoft C++. Can use other compilers (i.e. mingw32) if you have a separate Python installation.
• Cython >= 0.17. This is optional, as .cpp files generated by Cython are provided

NOTES:

1. ABAQUS is a commerical software package and requires a license from Simulia
2. The authors of pyvXRAY are not associated with ABAQUS/Simulia
3. Python and numpy are heavily used by pyvXRAY. These are built in to ABAQUS. All of the last few releases (v6.11 - v6.13) use Python 2.6.x and numpy 1.4.x

Model setup requirements

• The model must contain only linear or quadrilateral tetrahedral elements (ABAQUS element types C3D4, C3D4H, C3D10, C3D10H, C3D10I, C3D10M, and C3D10MH are all supported).

• The model must have a scalar fieldoutput variable representing bone density. This is typically a state variable such as SDV1. This scalar variable must be available in the last frame of each step to be analysed, as only the last frame is used.

Installation

####1. Installation of pyvXRAY plug-in

pyvXRAY is an ABAQUS plug-in. ABAQUS plug-ins may be installed in several ways. Only one of the ways is discussed here. For other options the user is referred to the ABAQUS user manuals.

• Releases with pre-built modules

  • Download the latest pyvXRAY release with pre-built modules. This is available for 32-bit and 64-bit Windows from releases page

  • Unpack this to a convenient location

  • Move the abaqus_plugins\pyvXRAY folder to the correct location of the abaqus_plugins directory within your ABAQUS installation. The location of this directory depends on your ABAQUS version. Some possible locations are:

    v6.11-x: C:\SIMULIA\Abaqus\6.11-x\abaqus_plugins

    v6.12-x: C:\SIMULIA\Abaqus\6.12-x\code\python\lib\abaqus_plugins

    v6.13-x: C:\SIMULIA\Abaqus\6.13-x\code\python\lib\abaqus_plugins

• Installation from source

  • Download the latest pyvXRAY source, typically called pyvXRAY-x.x.x.zip or pyvXRAY-x.x.x.tar.gz, where x.x.x is the version number

  • Unpack this to a convenient location

  • Open a command prompt and browse to directory pyvXRAY-x.x.x (containing file setup.py)

  • Run the following command:

        abaqus python setup.py build_ext --inplace
    

    which will build the Cython modules. If Cython is available, it will be used. Otherwise the .cpp files previously generated using Cython will be compiled directly.

  • Copy the pyvXRAY sub-folder to the abaqus_plugins directory within your ABAQUS installation, following the instructions above for pre-built distribution

####2. Installation of pyvXRAY dependencies

The ABAQUS GUI is built on Python, and has its own Python installation. This Python installation is not the typically Python setup, so some guidance is provided here on how to install pyvXRAY's dependencies.

Currently pyvXRAY has only one dependency that is not part of the ABAQUS Python, which is PIL / Pillow (NOTE: Pillow is a PIL fork that appears to have largely superseded PIL). On Windows it is easiest to install PIL / Pillow using a binary installer, particularly because PIL / Pillow have many dependencies. There are currently several issues with this:

• ABAQUS Python is typically not registered in the Windows registry, and therefore installation with binary installers will not work by default because the ABAQUS Python installation does not appear in the list of available Python installations

• PIL binaries are available only for Python 2.6 on 32-bit Windows at pythonware.com, but not 64-bit Windows. However, Pillow binaries for Python 2.6 on both 32-bit and 64-bit Windows are available on PyPi.

Given these limitations, there are two obvious choices for installating PIL / Pillow with ABAQUS Python.

• Use a separate Python installation and binary installer

  The SIMULIA support site suggests that a separate Python installation be setup. The dependencies can be installed easily into this separate Python installation, after which they can then be used by ABAQUS Python. This separate Python installation version must match the ABAQUS Python version, which has been 2.6.x for the following few ABAQUS versions. This method is the best solution if you do not feel comfortable modifying the Windows registry (as recommended next).

  • Download and install Python 2.6 (i.e. the plain vanilla version from www.python.org/download is recommended)

  • Download and run the corresponding binary installer for PIL or Pillow.

  • Create an environment variable PYTHONPATH=C:\Python26\Lib\site-packages that tells ABAQUS Python where these packages are installed, assuming that C:\Python26 is the installation directory.

• Edit the Windows registry and use a binary installer (Windows only)

  By editing the Windows registry the binary installers will be able to detect the ABAQUS Python version and install as usual. Use caution when editing the Windows registry or backup your registry before hand.

  • Download and run the corresponding binary installer for PIL or Pillow.

  • Edit the Windows registry to create key HKEY_LOCAL_MACHINE\Software\Python\Pythoncore\2.6\InstallPath with data name "(Default)" and data value containing the location of your ABAQUS Python directory location. Registry key HKEY_CURRENT_USER also works. This location depends on the ABAQUS version. For the default ABAQUS installation location, possible locations are:

    v6.11-x: C:\\SIMULIA\\Abaqus\\6.11-x\\External\\Python

    v6.12-x: C:\\SIMULIA\\Abaqus\\6.12-x\\tools\\SMApy

    v6.13-x: C:\\SIMULIA\\Abaqus\\6.13-x\\tools\\SMApy\\python2.6

    Editing the Windows registry can be done using the regedit utility. You can load regedit by typing "regedit" at the command prompt.

  • Install PIL / Pillow using the binary installer. Follow the instructions and make sure to select the ABAQUS Python version if you have multiple Python versions installed. If ABAQUS Python is not in the list of available Python 2.6 versions, then the Windows registry was not edited correctly.

Usage

• Open ABAQUS/CAE

• Open an odb file

• To launch the pyvXRAY GUI, go to the menubar at the top of the screen and select:

    Plug-ins --> pyvXRAY --> Create virtual x-rays
  

• Complete the required inputs in the GUI to suit the current model. More information is given below about the inputs

• Click OK to run pyvXRAY

• Look at the message area at the bottom of the screen for messages. On completion 'Finished' will be shown.

Required inputs

A basic description of each of the inputs required by pyvXRAY is listed here.

GUI tab	Input name	Input description
Select regions	Result file: Odb	The ABAQUS result file
	Bone region: Bone set	The name of the element set representing the bone
	Bone region: Density variable	A scalar fieldoutput variable representing bone density. This is most often a state variable i.e. SDV1
	Implant region: Show implant on x-rays	Option to include implant on the virtual x-rays
	Implant region: Implant set	The name of the element set representing the implant
	Implant region: Density (kg/m^3)	The density of the implant material in kg/m^3 i.e. 4500 for Titanium Alloy
Inputs	Required inputs: Step list	A list of steps to be analysed i.e. 1, 2, 3. A virtual x-ray is created for the last frame of each step in this list.
	Required inputs: Coordinate system	The name of the coordinate system used to create the projections. By default this is the global coordinate system. However, the views can be changed by creating a new coordinate system in ABAQUS and using it instead.
	Required inputs: Mapping resolution (mm)	pyvXRAY works by mapping the results of the bone density variable onto a regular grid. The mapping resolution is the cell spacing of this regular grid. Decreasing this number increases the accuracy of the mapping, but also increases the calculation time. As a first pass, a value of around 2mm is recommended to ensure that output is as expected.
X-ray settings	Settings: Base name of xray file(s)	This is the base or root name of the virtual x-ray image files. That is, image files are labelled basename_projection_stepnumber i.e. basename_XY_1 for the X-Y projection from Step 1.
	Settings: Approx size of x-ray images	Resizing of images is performed to make the number of pixels along the largest image dimension equal to this value.
	Settings: Image file format	Output format of images. Options are bmp, jpeg and png.
	Settings: Smooth images	Turn on image smoothing. PIL.ImageFilter.SMOOTH is used to perform the smoothing.
	Settings: Manual scaling of images	pyvXRAY scales the mapped bone density values when creating the virtual x-ray images. The image files are 24-bit (or 8-bit per channel), so the gray scale range is essentially 0-255. The scale factor used ensures that this range is fully utilised and that none of the images in the series are over-exposed. Activating this option reports the scale factors used and gives the user the ability to change these values. This may be desirable when comparing virtual x-rays from different models; an equal comparison is possible only if the same scale factors are used for both.

Outputs

pyvXRAY outputs a series of virtual x-rays correponding to the bone density results in a list of specified analysis steps. The bone density is mapped from the Finite Element Model to a overlapping regular grid of points and then projected onto each of the three Cartesian coordinate planes. If the model has an implant, then this can also be shown. The virtual x-ray images are saved in common image formats (bmp, jpeg, and png) and can be opened in any graphics package. These images can then be analysed to determine changes in the grey scale values, which can be related to the change in Bone Mineral Density (BMD) over time.

The recommended package for analysing these images is BMDanalyse, which is available free under the MIT license. BMDanalyse can be used to create regions of interest (ROIs) and determine the change in the average grey scale value within each ROI for all images in the series.

Help

For help create an Issue or a Pull Request on Github.