Eidors-logo    

EIDORS: Electrical Impedance Tomography and Diffuse Optical Tomography Reconstruction Software

EIDORS (mirror)
Main
Documentation
Tutorials
− Image Reconst
− Data Structures
− Applications
− FEM Modelling
− GREIT
− Old tutorials
Workshop
Download
Contrib Data
GREIT
Browse Docs
Browse SVN

News
Mailing list
(archive)
FAQ
Developer
                       

 

Hosted by
SourceForge.net Logo

 

Using Netgen to simulate fwd and inverse solutions

EIDORS can use Netgen to create sophisticated 2D and 3D models

Model a 2D tank with 8 electrodes

% Netgen simulation $Id: netgen_sims01.m 3273 2012-06-30 18:00:35Z aadler $

extra={'ball','solid ball = cylinder(0.2,0.2,0;0.2,0.2,1;0.2) and orthobrick(-1,-1,0;1,1,0.05);'};
fmdl= ng_mk_cyl_models(0,[9],[0.2,0,0.05],extra); 
ctr = interp_mesh(fmdl); ctr=(ctr(:,1)-0.2).^2 + (ctr(:,2)-0.2).^2;
img= mk_image(fmdl, 1 + 0.1*(ctr < 0.2^2));

subplot(221)
show_fem(img);

print_convert netgen_sims01a.png '-density 100'

Simulate a 2D tank with 8 electrodes with adjacent patterns

% Netgen simulation $Id: netgen_sims02.m 3791 2013-04-04 15:48:25Z aadler $

% Calculate stimulation pattern adjacent
stim = mk_stim_patterns(9,1,[0,1],[0,1],{},1);
img.fwd_model.stimulation = stim;

% Get all voltages so we can plot it
img.fwd_solve.get_all_meas = 1;

% Homogeneous model
img.elem_data(:) = 1;

vh = fwd_solve(img);

% Show Voltage for stim pattern #1
imgn = rmfield(img,'elem_data');
imgn.node_data = vh.volt(:,1);

h1= subplot(221);
show_fem(imgn);

% Show Voltage for stim pattern #2
imgn = rmfield(img,'elem_data');
imgn.node_data = vh.volt(:,2);

h2= subplot(222);
show_fem(imgn);

imgn.calc_colours.cb_shrink_move = [0.5,0.8,-.02];
common_colourbar([h1,h2],imgn);
print_convert netgen_sims02a.png '-density 100'


Figure: Voltage pattens for two adjacent simulation patterns

Simulate a 2D tank with 8 electrodes with different current patterns

% Netgen simulation $Id: netgen_sims03.m 3792 2013-04-04 16:01:34Z aadler $
clf;subplot(221);

% Calculate stimulation pattern adjacent
img.fwd_model.stimulation = mk_stim_patterns(9,1,[0,1],[0,1],{},1);
vh = fwd_solve(img);
imgn.node_data = vh.volt(:,1);

show_fem(imgn); axis equal; axis tight;
print_convert netgen_sims03a.png '-density 100'

% Calculate stimulation pattern offset-2
img.fwd_model.stimulation = mk_stim_patterns(9,1,[0,2],[0,1],{},1);
vh = fwd_solve(img);
imgn.node_data = vh.volt(:,1);

show_fem(imgn); axis equal; axis tight;
print_convert netgen_sims03b.png '-density 100'

% Calculate stimulation pattern offset-3
img.fwd_model.stimulation = mk_stim_patterns(9,1,[0,3],[0,1],{},1);
vh = fwd_solve(img);
imgn.node_data = vh.volt(:,1);

show_fem(imgn); axis equal; axis tight;
print_convert netgen_sims03c.png '-density 100'

% Calculate stimulation pattern offset-4
img.fwd_model.stimulation = mk_stim_patterns(9,1,[0,4],[0,1],{},1);
vh = fwd_solve(img);
imgn.node_data = vh.volt(:,1);

show_fem(imgn); axis equal; axis tight;
print_convert netgen_sims03d.png '-density 100'


Figure: Voltage pattens for two adjacent simulation patterns

Simulate a voltages for a change in conductivity

% Netgen simulation $Id: netgen_sims04.m 3792 2013-04-04 16:01:34Z aadler $

% Calculate stimulation pattern offset-4
img.fwd_model.stimulation = mk_stim_patterns(9,1,[0,4],[0,1],{},1);

% Set homogeneous phantom
img.elem_data(:) = 1;
vh = fwd_solve(img);
imgn.node_data = vh.volt(:,1);
imgn.calc_colours.cb_shrink_move = [0.3,0.7,0.02];
imgn.calc_colours.ref_level = 0;

clf; subplot(221);
show_fem(imgn,1); axis equal; axis tight;
print_convert netgen_sims04a.png '-density 100'

% Set inhomogeneity very insulating
img.elem_data = 1 - 0.99*(ctr < 0.2^2);
vi = fwd_solve(img);
imgn.node_data = vi.volt(:,1);

clf; subplot(221);
show_fem(imgn,1); axis equal; axis tight;
print_convert netgen_sims04b.png '-density 100'

% Difference in voltages
imgn.node_data = vh.volt(:,1) - vi.volt(:,1);

clf; subplot(221);
show_fem(imgn,1); axis equal; axis tight;
print_convert netgen_sims04c.png '-density 100'


Figure: Voltage pattens without (left) and with (middle) non-conductive target. Right is the difference in voltage pattern.

Voltage patterns for a change in conductivity

% Netgen simulation $Id: netgen_sims05.m 2787 2011-07-14 21:52:22Z bgrychtol $

subplot(211)
plot([vh.meas, vi.meas, vi.meas-vh.meas]);
print_convert netgen_sims05a.png '-density 100'


Figure: Voltage pattens without (green) and with (blue) non-conductive target. Red is the difference in voltage pattern.

Simple image reconstruction

% Netgen simulation $Id: netgen_sims06.m 4839 2015-03-30 07:44:50Z aadler $
subplot(221)

% Get a basic inverse model - replace the fwd model part
imdl = mk_common_model('a2c2',16); % the parameters aren't important because we replace it
imdl.fwd_model = img.fwd_model;

imgr = inv_solve(imdl, vh, vi);
show_fem(imgr);
axis equal; axis tight; print_convert netgen_sims06a.png '-density 100'

% Change the hyperparameter
imdl.hyperparameter.value = .05;
imgr = inv_solve(imdl, vh, vi);
show_fem(imgr);
axis equal; axis tight; print_convert netgen_sims06b.png '-density 100'


Figure: Image reconstructed with large (left) and smaller (right) hyperparameter value. Note that the expected resolution with nine electrodes isn't particularly good.

Last Modified: $Date: 2017-02-28 13:12:08 -0500 (Tue, 28 Feb 2017) $ by $Author: aadler $