Run a registration of a series of 2d images.
mia-2dmyoica-nonrigid2 -i <in-file> -o <out-file> [options]
mia-2dmyoica-nonrigid2 This program runs the non-rigid registration of an perfusion image series.In each pass, first an ICA analysis is run to estimate and eliminate the periodic movement and create reference images with intensities similar to the corresponding original image. Then non-rigid registration is run using the an "ssd + divcurl" cost model. The B-spline c-rate and the divcurl cost weight are changed in each pass according to given parameters.In the first pass a bounding box around the LV myocardium may be extractedto speed up computation Special note to this implemnentation: the registration is always run from the original images to avoid the accumulation of interpolation errors.
File-IO
input perfusion data set
output perfusion data set
file name base for registered fiels
save cropped set to this file
save segmentation feature images and initial ICA mixing matrix
ICA
ICA components 0 = automatic estimation
don't normalized ICs
don't strip the mean from the mixing curves
segment and scale the crop box around the LV (0=no segmentation)
skip images at the beginning of the series e.g. because as they are of other modalities
maximum number of iterations in ICA
Segmentation method
delta-peak \(hy difference of the peak enhancement images
features \(hy feature images
delta-feature \(hy difference of the feature images
Registration
Optimizer used for minimization For supported plugins see PLUGINS:minimizer/singlecost
start coefficinet rate in spines, gets divided by --c-rate-divider with every pass
cofficient rate divider for each pass
start divcurl weight, gets divided by --divcurl-divider with every pass
divcurl weight scaling with each new pass
image cost weight
image interpolator kernel For supported plugins see PLUGINS:1d/splinekernel
multi-resolution levels
registration passes
verbosity of output, print messages of given level and higher priorities. Supported priorities starting at lowest level are:
info \(hy Low level messages
trace \(hy Function call trace
fail \(hy Report test failures
warning \(hy Warnings
error \(hy Report errors
debug \(hy Debug output
message \(hy Normal messages
fatal \(hy Report only fatal errors
print copyright information
print this help
print a short help
print the version number and exit
Maxiumum number of threads to use for processing,This number should be lower or equal to the number of logical processor cores in the machine. (-1: automatic estimation).
bspline
B-spline kernel creation , supported parameters are:
d = 3 (int)
Spline degree. in [0, 5]
omoms
OMoms-spline kernel creation, supported parameters are:
d = 3 (int)
Spline degree. in [3, 3]
gdas
Gradient descent with automatic step size correction., supported parameters are:
ftolr = 0 (double)
Stop if the relative change of the criterion is below.. in [0, INF]
max-step = 2 (double)
Minimal absolute step size. in [1, INF]
maxiter = 200 (uint)
Stopping criterion: the maximum number of iterations. in [1, 2147483647]
min-step = 0.1 (double)
Maximal absolute step size. in [1e-10, INF]
xtola = 0.01 (double)
Stop if the inf-norm of the change applied to x is below this value.. in [0, INF]
gdsq
Gradient descent with quadratic step estimation, supported parameters are:
ftolr = 0 (double)
Stop if the relative change of the criterion is below.. in [0, INF]
gtola = 0 (double)
Stop if the inf-norm of the gradient is below this value.. in [0, INF]
maxiter = 100 (uint)
Stopping criterion: the maximum number of iterations. in [1, 2147483647]
scale = 2 (double)
Fallback fixed step size scaling. in [1, INF]
step = 0.1 (double)
Initial step size. in [0, INF]
xtola = 0 (double)
Stop if the inf-norm of x-update is below this value.. in [0, INF]
gsl
optimizer plugin based on the multimin optimizers ofthe GNU Scientific Library (GSL) https://www.gnu.org/software/gsl/, supported parameters are:
eps = 0.01 (double)
gradient based optimizers: stop when |grad| < eps, simplex: stop when simplex size < eps.. in [1e-10, 10]
iter = 100 (int)
maximum number of iterations. in [1, 2147483647]
opt = gd (dict)
Specific optimizer to be used.. Supported values are:
bfgs \(hy Broyden-Fletcher-Goldfarb-Shann
bfgs2 \(hy Broyden-Fletcher-Goldfarb-Shann (most efficient version)
cg-fr \(hy Flecher-Reeves conjugate gradient algorithm
gd \(hy Gradient descent.
simplex \(hy Simplex algorithm of Nelder and Mead
cg-pr \(hy Polak-Ribiere conjugate gradient algorithm
step = 0.001 (double)
initial step size. in [0, 10]
tol = 0.1 (double)
some tolerance parameter. in [0.001, 10]
nlopt
Minimizer algorithms using the NLOPT library, for a description of the optimizers please see 'http://ab-initio.mit.edu/wiki/index.php/NLopt_Algorithms', supported parameters are:
ftola = 0 (double)
Stopping criterion: the absolute change of the objective value is below this value. in [0, INF]
ftolr = 0 (double)
Stopping criterion: the relative change of the objective value is below this value. in [0, INF]
higher = inf (double)
Higher boundary (equal for all parameters). in [INF, INF]
local-opt = none (dict)
local minimization algorithm that may be required for the main minimization algorithm.. Supported values are:
gn-orig-direct-l \(hy Dividing Rectangles (original implementation, locally biased)
gn-direct-l-noscal \(hy Dividing Rectangles (unscaled, locally biased)
gn-isres \(hy Improved Stochastic Ranking Evolution Strategy
ld-tnewton \(hy Truncated Newton
gn-direct-l-rand \(hy Dividing Rectangles (locally biased, randomized)
ln-newuoa \(hy Derivative-free Unconstrained Optimization by Iteratively Constructed Quadratic Approximation
gn-direct-l-rand-noscale \(hy Dividing Rectangles (unscaled, locally biased, randomized)
gn-orig-direct \(hy Dividing Rectangles (original implementation)
ld-tnewton-precond \(hy Preconditioned Truncated Newton
ld-tnewton-restart \(hy Truncated Newton with steepest-descent restarting
gn-direct \(hy Dividing Rectangles
ln-neldermead \(hy Nelder-Mead simplex algorithm
ln-cobyla \(hy Constrained Optimization BY Linear Approximation
gn-crs2-lm \(hy Controlled Random Search with Local Mutation
ld-var2 \(hy Shifted Limited-Memory Variable-Metric, Rank 2
ld-var1 \(hy Shifted Limited-Memory Variable-Metric, Rank 1
ld-mma \(hy Method of Moving Asymptotes
ld-lbfgs-nocedal \(hy None
ld-lbfgs \(hy Low-storage BFGS
gn-direct-l \(hy Dividing Rectangles (locally biased)
none \(hy don't specify algorithm
ln-bobyqa \(hy Derivative-free Bound-constrained Optimization
ln-sbplx \(hy Subplex variant of Nelder-Mead
ln-newuoa-bound \(hy Derivative-free Bound-constrained Optimization by Iteratively Constructed Quadratic Approximation
ln-praxis \(hy Gradient-free Local Optimization via the Principal-Axis Method
gn-direct-noscal \(hy Dividing Rectangles (unscaled)
ld-tnewton-precond-restart \(hy Preconditioned Truncated Newton with steepest-descent restarting
lower = -inf (double)
Lower boundary (equal for all parameters). in [INF, INF]
maxiter = 100 (int)
Stopping criterion: the maximum number of iterations. in [1, 2147483647]
opt = ld-lbfgs (dict)
main minimization algorithm. Supported values are:
gn-orig-direct-l \(hy Dividing Rectangles (original implementation, locally biased)
g-mlsl-lds \(hy Multi-Level Single-Linkage (low-discrepancy-sequence, require local gradient based optimization and bounds)
gn-direct-l-noscal \(hy Dividing Rectangles (unscaled, locally biased)
gn-isres \(hy Improved Stochastic Ranking Evolution Strategy
ld-tnewton \(hy Truncated Newton
gn-direct-l-rand \(hy Dividing Rectangles (locally biased, randomized)
ln-newuoa \(hy Derivative-free Unconstrained Optimization by Iteratively Constructed Quadratic Approximation
gn-direct-l-rand-noscale \(hy Dividing Rectangles (unscaled, locally biased, randomized)
gn-orig-direct \(hy Dividing Rectangles (original implementation)
ld-tnewton-precond \(hy Preconditioned Truncated Newton
ld-tnewton-restart \(hy Truncated Newton with steepest-descent restarting
gn-direct \(hy Dividing Rectangles
auglag-eq \(hy Augmented Lagrangian algorithm with equality constraints only
ln-neldermead \(hy Nelder-Mead simplex algorithm
ln-cobyla \(hy Constrained Optimization BY Linear Approximation
gn-crs2-lm \(hy Controlled Random Search with Local Mutation
ld-var2 \(hy Shifted Limited-Memory Variable-Metric, Rank 2
ld-var1 \(hy Shifted Limited-Memory Variable-Metric, Rank 1
ld-mma \(hy Method of Moving Asymptotes
ld-lbfgs-nocedal \(hy None
g-mlsl \(hy Multi-Level Single-Linkage (require local optimization and bounds)
ld-lbfgs \(hy Low-storage BFGS
gn-direct-l \(hy Dividing Rectangles (locally biased)
ln-bobyqa \(hy Derivative-free Bound-constrained Optimization
ln-sbplx \(hy Subplex variant of Nelder-Mead
ln-newuoa-bound \(hy Derivative-free Bound-constrained Optimization by Iteratively Constructed Quadratic Approximation
auglag \(hy Augmented Lagrangian algorithm
ln-praxis \(hy Gradient-free Local Optimization via the Principal-Axis Method
gn-direct-noscal \(hy Dividing Rectangles (unscaled)
ld-tnewton-precond-restart \(hy Preconditioned Truncated Newton with steepest-descent restarting
ld-slsqp \(hy Sequential Least-Squares Quadratic Programming
step = 0 (double)
Initial step size for gradient free methods. in [0, INF]
stop = -inf (double)
Stopping criterion: function value falls below this value. in [INF, INF]
xtola = 0 (double)
Stopping criterion: the absolute change of all x-values is below this value. in [0, INF]
xtolr = 0 (double)
Stopping criterion: the relative change of all x-values is below this value. in [0, INF]
Register the perfusion series given in 'segment.set' by using automatic ICA estimation. Skip two images at the beginning and otherwiese use the default parameters. Store the result in 'registered.set'. mia-2dmyoica-nonrigid2 -i segment.set -o registered.set -k 2
Gert Wollny
This software is Copyright (c) 1999\(hy2013 Leipzig, Germany and Madrid, Spain. It comes with ABSOLUTELY NO WARRANTY and you may redistribute it under the terms of the GNU GENERAL PUBLIC LICENSE Version 3 (or later). For more information run the program with the option '--copyright'.