CIBC:Meetings:Management:Agenda
From NCRR Biomedical Software Development, Engineering, and Dissemination Wiki
Contents |
Continuous Collaborator/Developer Agenda
Goal of this page is to keep an on going list of projects needs, and accomplishments for each project and a short update of what the developers are up to.
Triedman Collaboration
Recent progress
a) We have a hexahedral approach working and we have surfaces for two of the segmented models.
b) We have a list of locations to put the electrodes and scans of electrode cans and wires (all specifications)
c) We have thresholds and voltage specifications.
d) We have code to do intersections in surfaces
To do
a) Test hexahedral approach, test different densities of meshes, test different ways of averaging conductivities.
b) Test efficiency in estimating impedance (one of the few measures we have we can use for validating the models, the ICD actually measures this :) )
c) Build electrode models from scans (probably manually because it is a 2D scan of 3D object and we need to add a depth)
d) Test different refinements around electrodes
e) Test whether insertion of electrode cans (closed surfaces) in surface models work.
f) Test effectiveness of mesh smoothing on results.
g) Generate a few more surface models with different parameters.
h) Use more elements in models (large dependency on element size in first tests) and hence fix memory management
Collaborator
a) Generate a few more segmented models and help with conversion mesh effectiveness study.
Development to do for developers
a) Fix problems intersection code
b) Expose more options in mesh smoother (currently no control over process)
c) Render a decent interface to TetGen
d) Fix MeshQuality module both GUI as code (crashes often and only displays data in shell)
e) Adjust intersection code to generate line electrodes at intersection of two planes.
f) FIX PROBLEMS IN SCIRUN SCHEDULER/INTERFACE
g) FIX MEMORY ADDRESS SPACE, BY STREAMING DATAFLOW TO DISK
h) Merge Jason's code (mesh smoother) into the branch
Short term goals
a) Upgrade modules to use consistent gets and sends for dataflow data
b) Create a system to cache data on disk
c) Fix interwsection doe and surface smoother
Henriquez Collaboration
Recent Progress
a) Fixed most bugs in the modules that generate the model
b) Add support for mapping electrodes in the model (we were getting data files that are too big > 100Gb) and hence we only will record at certain locations in high detail.
c) Small models seem to be produced OK.
d) Upgraded CardioWave code to record highly detailed electrodes
To do
a) Do simulations and write paper for CIC
b) Fix the bugs in latest version of CardioWave
For collaborative project
a) Time dependent electrodes in CardioWave (SCIRun side is done)
Collaborator
a) Work on paper
Development to do for developers
a) FIX SCIRUN GUI/SCHEDULER (So collaborator can use SCIRun)
b) Wrap modules so they are exposed in SCIRun Core (We currently though have a working set)
Short term goal
a) CIC paper has been written and has been accepted
Makeig/Worrell Collaboration
Recent Progress
a) We had a meeting with Greg Worrell and Scott Makeig:
1) They will send us CT data (better for segmentation skull)
2) We will make surface meshes using afront
3) We will try to minimize number of elements, they have a BEM guy starting to work in Scott's lab soon
4) We will try to build a Tet model out of the surfaces
5) We will look at mesh quality
=> Waiting for data, need to page him before leaving to Spain for CIC (Jeroen)
To do
a) Figure out how to segment their data and get painter application working
b) Surface Remesher (LONG TERM, will take months to complete)
c) Fix the current insertion code (Dana has some requirements here as well, hence we want to properly integrate over surfaces etc, hence our developers will be lost at this one , as it requires a lot of functionality that is already in the basis classes)
d) Add general dipole model for each type of mesh.
Development to do for developers
a) Get the painter application working and start playing with data and depending on that implement segmentation models. (I need input here, as I am not an expert in segmentation)
Short term goal:
1) Get data from Greg
2) Get data to McKay for 1st try at segmenting skull using CT data
Ellisman Collaboration
Please see CIBC:Project:NCMIR
Capecchi Collaboration
Overview: We are generating data for 2 papers. The first will be a comparison with the original bone length study from the Capecchi lab using our semi-automated techniques. The second paper will be a shape study using our shape tools. We have several students working on the necessary segmentations. Josh is working on shape analysis library and research.
Current Hox D11 work
- Zach, Lance, and Toly are currently working on bone segmentations.
- Zach is breeding new mutants: we need approx 3-4 more mutant scans and 3-4 more wild type scans for a total of 17 each. This will replicate the number used in the original, manual length measurement paper. Estimated we can scan the new mutants in December.
- Toly has length measurement software up and running and processing segmentations as we get them.
- Preliminary results show good agreement in relative bone length of mutant vs. wild.
- Targeting first publication for early 2007: this paper will be a bone length study comparing
- May use this project to beta test the NCMIR segmentation app when manual features ready.
Current Shape Analysis work
- Josh is working on a particle system library of code for use in shape analysis and surface visualization.
- Josh/Ross/Tom pursuing research in new methods for shape analysis. Josh gave talk for Mathematical Foundations of Computational Anatomy Workshop at MICCAI 2006.
Notes 9/11/06 - 9/29/06 Josh cranks on particle system code & MICCAI Talk. Otherwise, continue breeding and segmenting. Possible wild-type scanning work.
10/01/06 - Particle system code nearing completion. Zach, Lance recruiting more segmenters. Zach preparing poster presenting preliminary work.
SCIRun Release
- We have new naming convention !
- CMake conversion is going slower than planned (issues with dynamic compilation and packages), this is still not resolved.
- We blocked all editing while running a network (Allowing editing and making networks run fine turns out to be a hugh task for the developers.) This way users should not be able to crash SCIRun this way.
There are issues here we subnetworks, and callbacks (With subnetworks execution completion is not notified, we need to resolve this)
- CMAKE is not fully functional, configure has been fixed (at least yesterday it was working).
- No progress yet on documentation (Issue here which Wiki to use)
- Latest compiler gcc 4.1 on OSX is this working. The 4.2 troubles seem be from a beta beta version of GCC. We are not following up on this until that version becomes a stable version.
- Can we start to make the splash ppm for new version of SCIRun.
- Do we postpone performance issues LinAlgebra to next release ? (Our current Matirx class are an absolute mess)
Developers
Marty
a) Creating tools for inactive mesh editing
b) Working on the module renaming
c) Will work on updating the Doxygen documentation (do we want this?)
d) Will work on making small how to make a module tutorial (do we want this?)
Michael
a) Helping Darby with CMake branch, (this needs to be done before we can do the big reorganization)
b) Helping McKay to get Lexov working (we are already pst deadline)
McKay
a) Lexov
b) CMake switch, to coordinate things needed for release Painter under CMake
Darby
a) CMake switch
Test
CIBC:Documentation:MatlabInterface:Interface:InterfaceToMatlab
