View Full Version : Responses to "MRI scan into 3D image"

06-09-2009, 06:10 AM
Thank you for all of the helpful replies regarding "MRI scan into 3D image." As requested, a summary of the responses follows.


Sarah Day
Department of Mechanical & Manufacturing Engineering
Miami University
Oxford, Ohio, USA
:::::::::::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::.........
:::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::::: :::::.....

There are many commercial softwares that can do this. You can try Amira, Analyze, MIMICS or Rhinoceros - NURBS modeling for windows - to name a few.

I have personally used Analyze, Rhino and MIMICS and both have capability to render 3D graphics.

Best luck,

I am a little confused by your question. An MRI scan should give you a 3D image (ie, point data on a 3D orthogonal grid). If you can tell me a little more about what you plan on doing with the images, I might be able to give you some more help (FEA modeling, visualization, registration, etc).
I know that there are several commercial software packages available that can do this. Have you looked into 3D Slicer (www.slicer.org). Slicer is an open source software package that we have been using for converting MRI to geometries and image analysis. It is relatively easy to use and the software is free. If you just need to convert MRI to 3D geometries (i.e. STL or VTK), 3D Slicer would be worth looking into.
Trent M. Guess Ph.D.
Assistant Professor, Mechanical Engineering
350K Robert H. Flarsheim Hall
University of Missouri-Kansas City
5100 Rockhill Road
Kansas City, MO 64110-2499
Phone: 816.235.1252
FAX: 816.235.1260
e-mail: guesstr@umkc.edu
website: http://www1.sce.umkc.edu/~guesstr/
You can use: MIMICS/3D doctor/etc. or other available medical image processing software.
Hope this help.
With best wishes,
Dr. M. L. Jain,
(Visiting Scholar, www.uiuc.edu, USA)

Have you tried MIMICS or one of their other products? I think 3-Matic is
another one of their conversion products
Lieselle Trinidad
Doctoral Candidate
Mechanical Engineering Department
E-Lab Building, Complex 207e
160 Governors Drive
Amherst, MA 01003

Do you want to just make images, or to create finite-element models?
I am developing my own software for this purpose:
I don't know if it's 'highly regarded', but it's free :-)
I also have some relevant information on my Web site:
- Robert

I use Osirix (Mac OS only). That works pretty well.

PhD Student
ISM E. J. Marey (CNRS - Université Marseille 2)
Faculté des Sciences du Sport, case 910
163, Avenue de Luminy, 13288 Marseille

Regarding your question "MRI scan into 3-D image" on the BIOMCH-L.
A former colleague of mine use MIMICS for converting MRI-scans to 3D-models.
You can try the free software (http://www.stanford.edu/~loening/manuscripts/Loening_2003_AMIDE_postprint.pdf)to reconstruct MR 3D image. If you have any question, please feel free to ask me.

************************************************** **
You-Yin Chen, Ph.D.
Department of Electrical Engineering,
National Chiao-Tung University
Hsinchu,Taiwan, ROC
Tel: +886-3-5712121 ext:54427
Fax: +886-3-6125059
Mobile Phone: +886-9-37820318
Email: irradiance@so-net.net.tw

I do quite a bit of this. There are a number of options that I am familiar with in no particular order of preference (I will give some of my thoughts at the end) Hope you don't mind the length. Good luck with your work! There are a number of packages that could do the job.

1. Simpleware:

Simpleware is a suite of programs ScanIP, ScanFE and ScanCAD. I have used ScanIP and ScanFE. As the names suggest ScanIP is the image processing part of the suite. ScanIP can import DICOM, raster images and raw files, and export some of these also. It allows multiple "masks" for different materials/volumes etc. There are both manual and automatic segmentation tools, and boolean operations to allow union, subtraction, intersection and inversion of masks. A number of operations are valid over the entire volume, a region or a local slice.

There are a number of background filters, ranging from gaussian to nonlinear edge preserving filters. Other filters and operations include morphological mask operations, and a number of different level set segmentation implementations. The surfaces/volumes that you define in ScanIP can be then exported to ScanFE. ScanFE exists to convert the surfaces defined in ScanIP to FE/CFD meshes using solid elements. The features that I find nice in this are 1.) material allocation based on underlying greyscale value and 2) native export of mesh to various solvers (eg I use ANSYS and this is the only tool that I use that exports a file in ANSYS' text format directly, otherwise I need to add another step).


AMIRA and AVIZO are similar products, they used to simply be AMIRA, but a year or so ago the program was rebundled and split into AMIRA for life sciences and AVIZO for simulation, materials science and several other things. This probably means that there is more flexibility in what you licence but if you need breadth of product there are more modules to buy. I am not sure of how AMIRA has changed since the split as the installation that I use is now on AVIZO. These are general purpose high end visualisation, reconstruction and analysis packages. There are all the usual segmentation procedures. There is a neat level set contour tool for active highlighting. One of the features that I find very helpful in AVIZO is the registration module to align different scan volumes. There is also a link between the packages and MATLAB so that you can pass information back and forth for example to incorporate a custom filter. You can import just about anything. There is
tcl scripting to automate repetitive tasks. There are good tools to downsample and clean up triangulated surfaces, this can be a little time consuming if you have intersections. Tet meshes can be generated from triangulated surfaces. In the version I use I needed to do an extra conversion step to import it into ANSYS, but this is not a real problem. These programs can also be used for presentation of data and postprocessing and diffusion tensor imaging (maybe only AMIRA for DTI).

3. 3D Slicer

3D slicer is a free open source program used for research. I have used it but not extensively. I believe that it incorporates the VTK as its basis. I may be wrong here, so others may fill you in on this better. There is a broad scope of modules incorporated to generate surfaces, volumes, it looks like there is a new meshing module, registration and alignment, DTI etc.

I have not had much experience with MIMICS but it is a commercial product by MATERIALISE. It is used for 3D reconstruction, rapid prototyping and FE mesh generation, and possibly other things as well. Worth checking out. I think it includes marterial allocation based on greyscale value

My thoughts.

All of these packages are powerful. I personally like AMIRA/AVIZO and Simpleware. Amira for the manual segmentation tools and general flexibility of the product, and Simpleware for some meshing features.

The general purpose nature of AMIRA and AVIZO are very powerful, it is a fully featured visualisation package. I have benefited from being able to script a number of repetitive tasks across a number of models. I find the user interface clean, although I feel that the help leaves a little to be desired. I have found that I can do pretty much what ever I want in it, but it might take me a while to work it out. Calculations etc can be performed on your data to give all manner of things. You are probably limited by your imagination.

Simpleware makes good meshes and some of the tools are nice such as contact region definition on volume boundaries which can save time downstream, and material property allocation. Some of the other people at our school like the direct link to COMSOL also.

3D Slicer appears to be powerful, but there will be others more experienced in it than me. I did have problems with it crashing heaps several releases ago on a particular operation I needed to do but I understand that bug is fixed now.

MIMICS also seems powerful, but I cannot say much more about it.
Andrew Sims
PhD Student
Graduate School of Biomedical Engineering
University of New South Wales
Sydney, Australia
+61 2 9385 39
You don't indicate what you want to do once you have a 3D rendering. Are you interested in just visualizing it or do you want to do some processing? We use VTK from Kitware Inc. http://vtk.org/ It is powerful but requires some programming experience. It's free.
Here are a few that I have investigated. It all depends on what structures you want to look at from your MRIs and how much you want to spend. All companies will offer you a free trial period for their software. Just contact them via their websites. Make sure that your MRI file format is easily loaded into the software. Most are compatible with DICOM files. Also check what output files are produced from the 3D software as you may want to export the 3D rendering into other programs and manipulate it.

Mimics - Expensive, but very user friendly.
3D Slicer/Slicer 2.6 - this is free!! It is a little clunky to use but does just about everything.
If I can assist in any other way please contact me.
Lee Barber
School of Physiotherapy and Exercise Science
Gold Coast Campus
Queensland 4222

Was your MRI scanned using the appropriate sequence to be reconstructed in 3D? If so, then I know Analyze will load and reconstruct the volume from the MR data. I believe Mimics will also do the same. It would help to know your ultimate goal with the images to know which of the two may be more advantageous. Let me konw if you have any more questions.

I'm taking a class now where we are using software my professor created called SuperSlicer to generate 3D images from MRI/CT scans (primarily for subsequent use in a finite-element model).

The professor's name is Todd Doehring. You can reach him at tcd28@drexel.edu.

David Jamison, IV
Ph.D. Student, Drexel University

We are using Amira for about 10 years and are very happy with it.
See www.amiravis.com
I think you can download it for a free trial.
Best wishes,

You can use Amira (Mercury Software) or maybe "paraview". The latter is a freeware based on VTK.
Best regards
There is a free Brazilian software called InVesalius that uses MRI images to build 3D models. You can find it here:
Unfortunately, it is in Portuguese.
I am not sure if it is suitable for your needs, but there is another free software called CTMRedit (http://www.isle.illinois.edu/pubs/1990s/hasegawa-johnson99embs.pdf), that is a matlab based software for segmenting and interpolating MRI images.
Kind regards,
Karine Sarro

You may want to check out MIMICS (http://www.materialise.com/materialise/view/en/92458-Mimics.html)
Hope this can help.
Ryo Minoguchi
Procter & Gamble Co.

Osirix is a wonderful application for this type of rendering and it is
also free. The downside is that it only runs on Mac.

Bo Foreman, PT, PhD
University of Utah

Haven't used myself but AMIRA seems to be pretty straight forward to use. You can download a trial version from their homepage.
Hope this helps.

I use SliceoMatic (Tomovision, Montreal, Canada) to segment my MR images and turn them into 3-D polygon surfaces. This program allows you to use several different semi automated techniques to segment your images (region growing, snakes, thresholding, etc.). It also allows you to save your solid model in a variety of formats (i.e. stl,mov,dxf,VRML) for processing. I'm not sure if it is the best software out there but it is user friendly and fairly easy to use. I would really appreciate it if you could send me a final list of suggested products that you come up with. Thanks.

PhD Candidate
Musculoskeletal Biomechanics Research Laboratory
Division of Biokinesiology and Physical Therapy
School of Dentistry
University of Southern California

Simpleware works, you can also look at Mimics and Amira as well for 3D reconstruction. But if you are only interested in viewing the 3D image there should be freeware softwares that can be used for this purpose, atleast I am sure for CT images, not sure for MRI. What are you using the 3D image for, if I may ask?

Are you looking to get a 3D model, or just visualize in 3D? For software that will let you view your MRI data in 3D, check out "VolView":

If you pay, you can output 3D geometry. However, "VolView" can be used as a viewer for free and is very powerful.

There is a small piece of free software called "Reconstruct" that can be used to generate 3D models from your MRI data:

"Reconstruct" requires you to outline the geometry of interest, then view the results as a 3D model.

I've also used Mimics with great success, but it is expensive:

I hope this helps. You might also search the archives as I know this question has been asked before. However, it is always good to ask again in case new technology has emerged...
Jeremy J. Bauer, Ph.D.
Hayes + Associates, Inc.
301 SW 4th Street, Suite 160
Corvallis, OR 97333

Materialise is another big name in this field, their software "Mimics" is probably all you need, for MRI slice editing to mesh conversion.
( http://www.materialise.com/materialise/view/en/92458-Mimics.html )

For cheap/free software, to simply display a 3D image of the scan volume you might want to explore any of the software listed here:
( http://www.rtstudents.com/pacs/free-dicom-viewers.htm )

Good luck,

Hin Chung
EngD Student
Doctoral Training Centre in Medical Devices
Bioengineering Unit
University of Strathclyde

"AMIRA" is able to convert DICOM-Images into a 3D-Image.
Best regards,
Philip Kleinau
-- University of applied sciences Giessen-Friedberg
Biomechanics Laboratory
Wiesenstrasse 14
35390 Giessen/Germany

We've used Mimics from materialize (www.materialize.com You can request a 30 day trial on the website) to create 3-D reconstructions of the spine using both CT and MR images. It can take a little manual work to get the images looking really good, but it's not bad and it's pretty easy to learn. It can also be a bit pricey depending on which functions you need, but they do give a 50% educational/non-profit discount if you ask them about it.
Let me know if I can give you any more information.
Good luck,
Susan M. Renner, PhD
Biomedical Engineer
Musculoskeletal Biomechanics Laboratory,
Edward Hines Jr. VA Hospital

Mimics by Materialise
Jenni Buckley, PhD
UCSF/SFGH Biomechanical Testing Facility

The Visualization ToolKit (VTK) will do it. There are wrappers that will allow you to use it in C++, TCL, and/or python. If your looking for something a little more user friendly I would try slicer. There is a tutorial on how to do exactly that on their website (http://www.slicer.org/).

MIMICS from Materialise is a great way to go. We use it in our lab to segment CT and MR data, and it works quite well for both. They are also very responsive to customer feedback and with customer support. They offer a free trial as well.
Have a good one,

Check out MIMICS as well.
Cheers, DLL
Danny L. Levine, Ph.D.
Principal Engineer
Zimmer, Inc.

we have been using Materialise/ MIMICS software since last year in order to convert from CT/MRI scan to 3D structure.
i have used Simplaware (only scan ip) evaluation version for 10 days.
we did simulation ( preoperative study) , structurual analysis via MIMIC's models.
to me , Tutorials of Simpleware software is not adequate to learn the program.
And i know Simplaware software was released last year.
we can take help from Materialise office to overcome complex problem.

I am using Slicer (http://www.slicer.org/) to view and segment MRI pictures.
It's freeware and at
http://openpaleo.blogspot.com/2008/12/3d-slicer-tutorial.html you can find a
good tutorial to get started with the basic functions.
Best regards from Germany,

you may check out Materialise Mimics.

Try this open source software to see if it meets your needs:

- Ed -
Edward Morra, MSME Manager of Computational Testing Services Orthopaedic Research Laboratories
Cleveland, Ohio 44114

We have used simpleware very briefly (as a student project) and are also
looking at alternatives. I would value the chance to look at what
suggestions are made to you! Could you copy me into any summary you are able
to pull together.

Amira, Materialise
Peter Ellison MEng PhD
Research Manager
Corin Ltd | The Corinium Centre | Cirencester | Gloucestershire | GL7 1YJ

I am doing similar work with CT scans and came across 3D-Doctor doing a similar search. It's really easy to use and will create a detailed model for a variety of different scans (CT, MRI, etc...) and it's FDA approved for this purpose. http://www.ablesw.com/3d-doctor/
They have a demo version on their website that's pretty good for playing around, you can upload any images that you want and it'll let you create a model, you just can't save anything in the demo version (take screen shots if you need any images ASAP). They offer an academic license at about $2300 USD. The only limitation is that the animations are a bit limited in what they can do and show (or at least I haven't figured out eactly how to make them 'pretty' yet). Regardless, it'll create an accurate 3D model, it is an easy-to-use interface, and you don't have to buy add-on packages like a variety of other similar softwares (ie. all functions are included in the main pcakage). Not sure of your exact application or use but I hope this information can be of some use to you.
With kind regards,
Spencer Bell
COOP Student, 4th Year Biological Engineering
Graeme Clark Neuroscience Institute
La Trobe University
Melbourne, Australia

I heard Caret work pretty well, which is free.
This may require your manual segmentation and generation of subject volume from the image, which could be done using customized program in Matlab (if you are interested, I could share some source files with you).
Jiang Yao
Postdoc Research Fellow
Department of Mechanical Engineering
University of Rochester

I use 3D-DOCTOR (www.ablesw.com/3d-doctor) for my (micro-CT) image stacks. It does what I want and was a lot less expensive than Simpleware.

Several of them in the market
My first choice (I have already worked with it): MIMICS
Others by mind:
3D Doctor (http://www.ablesw.com/3d-doctor/3ddoctor.html)
Slicer (http://www.slicer.org/) - free open source
Best regards
Pedro Talaia

Try www.vmtk.org, a free open-source program that have similar functions.
Good luck!

Try "MRIcro":
Good luck,
Nick Roach
PhD student, Electronic Engineering/Psychology
University of Birmingham,UK

I really like Osirix. It's quick, cheap and easy to use and the results (I think) are as good as the expensive products. It is certainly worth a go before you splash out a lot of money.
--Dr. Bill Sellers William.Sellers@manchester.ac.uk
Programme Director of Zoology Faculty of Life Sciences
The University of Manchester

Other than Simpleware, try Mimics and Amira. All have their pros and cons and should be able to let you have a demo version to let you choose.
Dr Jonathan Jeffers
Senior Research and Development Engineer

Try amira its good