[vtkusers] Job Opening at New Startup

[John Axerio-Cilies]

Hi All,

I am in charge of development for a brand new startup in Silicon Valley.
There are currently four of us on the team, and we are in need of a VTK guru
to help us develop our product. If you think you can work remotely over the
next couple weeks on a simple project, we would love to hear from you.
Please send me an email at axerio at stanford.edu if you are interested. If you
can work remotely full-time or are in the bay area, even better! If it turns
out that we are happy with the simple script you develop, you will have an
opportunity to join the team full-time.

Your job would be create a script to automatically execute the following 8
steps that will eventually become part of the back-end of our system. We
don't need an executable, just a simple script that calls the appropriate
functions (no visualization necessary). This would require the following
steps:

1) Read DICOM data: the function needs to be able to handle either a
location where all the DICOM images are stored, a zipped file, or a DICOM
multi image file. (vtkDICOMImageReader or similar)
2) Image Smoothing and Segmentation: A seed point would be placed 1cm in
front of the geometry of interest, and the desired volume would need to be
segmented. I have used the region growing segmentation in the past using a
HU range of (-1000 to -180). Potentially we could use something more
complicated using Level Set-Based filters.
3) Create 3D Model: Use vtkContourFilter or similar
4) Smooth 3D Model: Use vtkSmoothPolyDataFilter or similar
5) Fill Holes in 3D Model: Use vtkFillHolesFilter or similar. If there are
still holes after this operation, the file needs to be flagged automatically
for us to look at manually.
6) Clip Data: Use vtkClipPolyData or similar to intersect a closed box with
the 3D volume.
7) Save to .vtk Fomat: Use vtkWriter or similar
8) Save to .stl Format: Use vtkSTLWriter or similar

Optional Steps:

9) DICOM Image Visualization: These can be viewed in ParaView or
ParaViewWeb, removing any patient identifying tags.
10) ParaViewWeb Visualization: The ability to run the ParaView server in
batch to save a particular state, and visualize on the web (using the WebGL
State functionality of ParaViewWeb).

If you can give me a general sense of how many hours each step will take, as
well as your hourly rate this would be really appreciated. Otherwise, we can
negotiate on a fixed price basis. I have about 5 DICOM sets (I can get more
if needed) that we can use to test the script for robustness and error
handling. We don't need a release version or anything, just a prototype that
we can use to start collecting clinical data.

Regards,

John

_____________________________________
John Axerio-Cilies, PhD Candidate
Flow Physics and Computational Engineering
Mechanical Engineering, Stanford University
axerio at stanford.edu
650 391-7111
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20110720/e41c04c0/attachment.htm>