Top Secret Drum Corps @ Edinburgh Military Tattoo 2006
Why? Just play the video….o…my..god. Basically played it about 1000 times last 2 years, time to share it with the world!
As said in previous posts I had to complete the whole traject of creating an educational software application. As time was limited you had to keep in mind the purpose was not to create the most beautiful application but to learn about all the things which accompany creation of such an application. Two major things are ofcourse the design of the application and the evaluation of the application.
The application
The application I will present on the 7th of june is made to educate medical students in the subject of structural brain connectibity (SBC). As SBC cannot be seen in a human cadaver, and is hard to explain using only classical education methods as textbooks and literature slides, an interactive application could prove to be useful. As shown by Dale’s cone of experience we remember 90% of what we do, supporting the use of an interactive application. For a complete specification of the Design principles you will have to wait until I publish my final report
. It will explain all educational principles as well as the principles of Multimedia learning which are kept in mind while designing the application.
The application consists of three views. Each view contains multiple visualizations except for the first one. The first view contains an Isosurface rendering of the human brain. The second view contains a volume rendering of the human brain, a fibre visualization of the (water) diffusion data, and a MRI slice viewer. All coordinate systems are overlayed meaning that the location in 3-D space is the same location in each of the visualizations in the view. The last view contains also the volume, and MRI slicer, but exchanges the fibre visualization for a glyph visualization.
This way the student can use the knowledge of known views to learn and use the new type of visualizations (the fibre view and glyph view). As students tend to be lazy, questions are posed which have to be answered using the various views. In this way the student is ‘forced’ to use the visualizations and becomes familar with them.
The question sets can be constructed by the instructor. In order to do this, the instructor puts a series of questions in a textfile accompanied with the view which has to be used (volume, fibres, glyph, combination of all) and the points received when the question is answered correctly. In a normal case the instructor will let the student first answer questions about all views seperately and in the end ask questions in which views have to combined, much like the way math books work, first learn the individual formula’s and in the end combine those to solve more complicated questions.
All results are logged and shown as feedback to the student and the logfile is made available to the instructor to enable him or her to check the student progress of the group.
DeVIDE and prototyping
To find the three correct views I have used DeVIDE. DeVIDE, or the Delft Visualisation and Image processing Development Environment, is a cross-platform software framework for the rapid prototyping, testing and deployment of visualisation and image processing algorithms. I have used it to quickly generate multiple visualizations and check those with two medical students to find the three best visualizations to put in the educational application. As DeVIDE is also programmed in Python, AND is open-source, I could 1-on-1 transfer those pipelines to my own application running on PyQt4 + vtk.
I would advise everyone using vtk for development to at least have a quick peek at DeVIDE, as it will enable you to quickly prototype various visualization pipelines without bothering for the correct code. If needed, DeVIDE provides a CodeRunner module which can be used to put python code in.
Final application screencast
I have made a quick screencast demo-ing the application which is embedded below. Have a look!
That’s all folks!
Since our publication actually made it to VCBM2010 (see earlier posts) it is now listed at the TUDelft publications site for our department. Be sure to check this site as there are other interesting papers there as well including:
- “Dynamic Visualisation of Orbital Fat Deformation”
- P. J. Schaafsma, S. Schutte, H. J. Simonsz, F. H. Post, and C. P. Botha - “Combined Surface and Volume Processing for Fused Joint Segmentation”
- P. R. Krekel, E. R. Valstar, F. H. Post, P. M. Rozing, and C. P. Botha
Next to this, this week or the next we will probably hear from Universiteitsfond Delft if they will cover part of the costs of presenting our work in Leipzig, Germany where VCBM 2010 is held. Our department already tossed in a couple of hundreds so lets hope they close the last gap so we can almost go for free …. i mean, we are still students right?
Recently a new version of Ubuntu went public. Ubuntu 10.04 Lucid introduces some improvements and some eye candy. Of course my linux machine was updated to run this version over the previously installed 9.10 Karmic build. Everything went well and I was happy as a little boy being handed a popsicle by his mother on a hot summer day in July. Until I found out..
Happy as I was after like a week or two I decided to run my python application which suddenly decided to crash bigtime. The application is created for Educational Software, a course at the TUDelft. It is nearly finished and running smoothly in 9.10 Karmic. It uses VTK (edited), PyQT and vtkTeem and produces some visualizations of structural connectivity in the brain. But suddenly, in ubuntu 10.04, creating a mask with as input an vtkTeem.NRRDReader() output and VTIReader output (casted to unsigned char) produces the following error:
*** glibc detected *** python: double free or corruption (out): 0x00007f0984000990 ***
My first guess is that this only produced a warning in the previous eglibc and in the new version is (correctly) handled by throwing an error. Since ubuntu relies heavily on eglibc a simple fix is out of the question. Seems we have to either dive into the VTK source and check what is wrong or really hope I made a stupid mistake somewhere in my code
).
ubuntu 1 – 0 bastijn
Stay tuned!
Last friday I received a mail with the following contents:
“We are very pleased to inform you that your paper has been ACCEPTED
for the Eurographics Workshop on Visual Computing for Biology and
Medicine (EG VCBM) 2010.”
After a little celebration that our paper got accepted at the conference it was time to read the comments of the reviewers. Those weren’t that bad at all, of course we need to do some work but I think we can do a lot before the 1st of July. Lets hope

