.TH mia\-2dmyoica\-nonrigid2 1 "v2.4.7"  "USER COMMANDS"
.SH NAME
mia\-2dmyoica\-nonrigid2
\- Run a registration of a series of 2D images.
.SH SYNOPSIS
.B  mia\-2dmyoica\-nonrigid2 \-i <in\-file> \-o <out\-file> [options]
.SH DESCRIPTION
.B 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.
.SH OPTIONS
.SS 
File-IO
.RS
.IP "\-i \-\-in-file=(required, input); string"
input perfusion data set

.IP "\-o \-\-out-file=(required, output); string"
output perfusion data set

.IP "\-r \-\-registered=reg"
file name base for registered fiels

.IP "   \-\-save-cropped="
save cropped set to this file

.IP "   \-\-save-feature="
save segmentation feature images and initial ICA mixing matrix

.RE
.SS 
ICA
.RS
.IP "   \-\-fastica=internal"
FastICA implementationto be used
 For supported plugins see PLUGINS:fastica/implementation
.IP "\-C \-\-components=0"
ICA components 0 = automatic estimation

.IP "   \-\-normalize"
don't normalized ICs

.IP "   \-\-no-meanstrip"
don't strip the mean from the mixing curves

.IP "\-s \-\-segscale=0"
segment and scale the crop box around the LV (0=no segmentation)

.IP "\-k \-\-skip=0"
skip images at the beginning of the series e.g. because as they are of other modalities

.IP "\-m \-\-max-ica-iter=400"
maximum number of iterations in ICA

.IP "\-E \-\-segmethod=features"
Segmentation method

.RS 10
.I
delta\-feature
\(hy difference of the feature images
.RE
.RS 10
.I
delta\-peak
\(hy difference of the peak enhancement images
.RE
.RS 10
.I
features
\(hy feature images
.RE
.RE
.SS 
Registration
.RS
.IP "\-O \-\-optimizer=gsl:opt=gd,step=0.1"
Optimizer used for minimization
 For supported plugins see PLUGINS:minimizer/singlecost
.IP "\-a \-\-start-c-rate=32"
start coefficinet rate in spines, gets divided by \-\-c\-rate\-divider with every pass

.IP "   \-\-c-rate-divider=4"
cofficient rate divider for each pass

.IP "\-d \-\-start-divcurl=20"
start divcurl weight, gets divided by \-\-divcurl\-divider with every pass

.IP "   \-\-divcurl-divider=4"
divcurl weight scaling with each new pass

.IP "\-w \-\-imageweight=1"
image cost weight

.IP "\-p \-\-interpolator=bspline:d=3"
image interpolator kernel
 For supported plugins see PLUGINS:1d/splinekernel
.IP "\-l \-\-mg-levels=3"
multi\-resolution levels

.IP "\-P \-\-passes=3"
registration passes

.RE
.SS Help & Info
.RS
.IP "\-V \-\-verbose=warning"
verbosity of output, print messages of given level and higher priorities. Supported priorities starting at lowest level are:

.RS 10
.I
trace
\(hy Function call trace
.RE
.RS 10
.I
debug
\(hy Debug output
.RE
.RS 10
.I
info
\(hy Low level messages
.RE
.RS 10
.I
message
\(hy Normal messages
.RE
.RS 10
.I
warning
\(hy Warnings
.RE
.RS 10
.I
fail
\(hy Report test failures
.RE
.RS 10
.I
error
\(hy Report errors
.RE
.RS 10
.I
fatal
\(hy Report only fatal errors
.RE
.IP "   \-\-copyright"
print copyright information

.IP "\-h \-\-help"
print this help

.IP "\-? \-\-usage"
print a short help

.IP "   \-\-version"
print the version number and exit

.RE
.SS Processing
.RS
.IP "   \-\-threads=\-1"
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).

.RE
.SH PLUGINS: 1d/splinekernel
.TP 10
.B bspline
B-spline kernel creation , supported parameters are: 
.P
.RS 14
.I
d
= 3; int in [0, 5] 
.RS 2
Spline degree.

.RE
.RE
.TP 10
.B omoms
OMoms-spline kernel creation, supported parameters are: 
.P
.RS 14
.I
d
= 3; int in [3, 3] 
.RS 2
Spline degree.

.RE
.RE
.SH PLUGINS: fastica/implementation
.TP 10
.B internal
This is the MIA implementation of the FastICA algorithm.
.P
.RS 14
(no parameters)
.RE
.TP 10
.B itpp
This is the IT++ implementation of the FastICA algorithm.
.P
.RS 14
(no parameters)
.RE
.SH PLUGINS: minimizer/singlecost
.TP 10
.B gdas
Gradient descent with automatic step size correction., supported parameters are: 
.P
.RS 14
.I
ftolr
= 0; double in [0, inf) 
.RS 2
Stop if the relative change of the criterion is below..

.RE
.RE
.RS 14
.I
max-step
= 2; double in (0, inf) 
.RS 2
Maximal absolute step size.

.RE
.RE
.RS 14
.I
maxiter
= 200; uint in [1, inf) 
.RS 2
Stopping criterion: the maximum number of iterations.

.RE
.RE
.RS 14
.I
min-step
= 0.1; double in (0, inf) 
.RS 2
Minimal absolute step size.

.RE
.RE
.RS 14
.I
xtola
= 0.01; double in [0, inf) 
.RS 2
Stop if the inf\-norm of the change applied to x is below this value..

.RE
.RE
.TP 10
.B gdsq
Gradient descent with quadratic step estimation, supported parameters are: 
.P
.RS 14
.I
ftolr
= 0; double in [0, inf) 
.RS 2
Stop if the relative change of the criterion is below..

.RE
.RE
.RS 14
.I
gtola
= 0; double in [0, inf) 
.RS 2
Stop if the inf\-norm of the gradient is below this value..

.RE
.RE
.RS 14
.I
maxiter
= 100; uint in [1, inf) 
.RS 2
Stopping criterion: the maximum number of iterations.

.RE
.RE
.RS 14
.I
scale
= 2; double in (1, inf) 
.RS 2
Fallback fixed step size scaling.

.RE
.RE
.RS 14
.I
step
= 0.1; double in (0, inf) 
.RS 2
Initial step size.

.RE
.RE
.RS 14
.I
xtola
= 0; double in [0, inf) 
.RS 2
Stop if the inf\-norm of x\-update is below this value..

.RE
.RE
.TP 10
.B gsl
optimizer plugin based on the multimin optimizers of the GNU Scientific Library (GSL) https://www.gnu.org/software/gsl/, supported parameters are: 
.P
.RS 14
.I
eps
= 0.01; double in (0, inf) 
.RS 2
gradient based optimizers: stop when |grad| < eps, simplex: stop when simplex size < eps..

.RE
.RE
.RS 14
.I
iter
= 100; uint in [1, inf) 
.RS 2
maximum number of iterations.

.RE
.RE
.RS 14
.I
opt
= gd; dict 
.RS 2
Specific optimizer to be used..
Supported values are:
.RS 4
.I
simplex
\(hy Simplex algorithm of Nelder and Mead
.RE
.RS 4
.I
cg\-fr
\(hy Flecher-Reeves conjugate gradient algorithm
.RE
.RS 4
.I
cg\-pr
\(hy Polak-Ribiere conjugate gradient algorithm
.RE
.RS 4
.I
bfgs
\(hy Broyden-Fletcher-Goldfarb-Shann
.RE
.RS 4
.I
bfgs2
\(hy Broyden-Fletcher-Goldfarb-Shann (most efficient version)
.RE
.RS 4
.I
gd
\(hy Gradient descent.
.RE

.RE
.RE
.RS 14
.I
step
= 0.001; double in (0, inf) 
.RS 2
initial step size.

.RE
.RE
.RS 14
.I
tol
= 0.1; double in (0, inf) 
.RS 2
some tolerance parameter.

.RE
.RE
.TP 10
.B 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: 
.P
.RS 14
.I
ftola
= 0; double in [0, inf) 
.RS 2
Stopping criterion: the absolute change of the objective value is below  this value.

.RE
.RE
.RS 14
.I
ftolr
= 0; double in [0, inf) 
.RS 2
Stopping criterion: the relative change of the objective value is below  this value.

.RE
.RE
.RS 14
.I
higher
= inf; double 
.RS 2
Higher boundary (equal for all parameters).

.RE
.RE
.RS 14
.I
local-opt
= none; dict 
.RS 2
local minimization algorithm that may be required for the main minimization algorithm..
Supported values are:
.RS 4
.I
gn\-direct
\(hy Dividing Rectangles
.RE
.RS 4
.I
gn\-direct\-l
\(hy Dividing Rectangles (locally biased)
.RE
.RS 4
.I
gn\-direct\-l\-rand
\(hy Dividing Rectangles (locally biased, randomized)
.RE
.RS 4
.I
gn\-direct\-noscal
\(hy Dividing Rectangles (unscaled)
.RE
.RS 4
.I
gn\-direct\-l\-noscal
\(hy Dividing Rectangles (unscaled, locally biased)
.RE
.RS 4
.I
gn\-direct\-l\-rand\-noscale
\(hy Dividing Rectangles (unscaled, locally biased, randomized)
.RE
.RS 4
.I
gn\-orig\-direct
\(hy Dividing Rectangles (original implementation)
.RE
.RS 4
.I
gn\-orig\-direct\-l
\(hy Dividing Rectangles (original implementation, locally biased)
.RE
.RS 4
.I
ld\-lbfgs\-nocedal
\(hy None
.RE
.RS 4
.I
ld\-lbfgs
\(hy Low-storage BFGS
.RE
.RS 4
.I
ln\-praxis
\(hy Gradient-free Local Optimization via the Principal-Axis Method
.RE
.RS 4
.I
ld\-var1
\(hy Shifted Limited-Memory Variable-Metric, Rank 1
.RE
.RS 4
.I
ld\-var2
\(hy Shifted Limited-Memory Variable-Metric, Rank 2
.RE
.RS 4
.I
ld\-tnewton
\(hy Truncated Newton
.RE
.RS 4
.I
ld\-tnewton\-restart
\(hy Truncated Newton with steepest-descent restarting
.RE
.RS 4
.I
ld\-tnewton\-precond
\(hy Preconditioned Truncated Newton
.RE
.RS 4
.I
ld\-tnewton\-precond\-restart
\(hy Preconditioned Truncated Newton with steepest-descent restarting
.RE
.RS 4
.I
gn\-crs2\-lm
\(hy Controlled Random Search with Local Mutation
.RE
.RS 4
.I
ld\-mma
\(hy Method of Moving Asymptotes
.RE
.RS 4
.I
ln\-cobyla
\(hy Constrained Optimization BY Linear Approximation
.RE
.RS 4
.I
ln\-newuoa
\(hy Derivative-free Unconstrained Optimization by Iteratively Constructed Quadratic Approximation
.RE
.RS 4
.I
ln\-newuoa\-bound
\(hy Derivative-free Bound-constrained Optimization by Iteratively Constructed Quadratic Approximation
.RE
.RS 4
.I
ln\-neldermead
\(hy Nelder-Mead simplex algorithm
.RE
.RS 4
.I
ln\-sbplx
\(hy Subplex variant of Nelder-Mead
.RE
.RS 4
.I
ln\-bobyqa
\(hy Derivative-free Bound-constrained Optimization
.RE
.RS 4
.I
gn\-isres
\(hy Improved Stochastic Ranking Evolution Strategy
.RE
.RS 4
.I
none
\(hy don't specify algorithm
.RE

.RE
.RE
.RS 14
.I
lower
= \-inf; double 
.RS 2
Lower boundary (equal for all parameters).

.RE
.RE
.RS 14
.I
maxiter
= 100; int in [1, inf) 
.RS 2
Stopping criterion: the maximum number of iterations.

.RE
.RE
.RS 14
.I
opt
= ld\-lbfgs; dict 
.RS 2
main minimization algorithm.
Supported values are:
.RS 4
.I
gn\-direct
\(hy Dividing Rectangles
.RE
.RS 4
.I
gn\-direct\-l
\(hy Dividing Rectangles (locally biased)
.RE
.RS 4
.I
gn\-direct\-l\-rand
\(hy Dividing Rectangles (locally biased, randomized)
.RE
.RS 4
.I
gn\-direct\-noscal
\(hy Dividing Rectangles (unscaled)
.RE
.RS 4
.I
gn\-direct\-l\-noscal
\(hy Dividing Rectangles (unscaled, locally biased)
.RE
.RS 4
.I
gn\-direct\-l\-rand\-noscale
\(hy Dividing Rectangles (unscaled, locally biased, randomized)
.RE
.RS 4
.I
gn\-orig\-direct
\(hy Dividing Rectangles (original implementation)
.RE
.RS 4
.I
gn\-orig\-direct\-l
\(hy Dividing Rectangles (original implementation, locally biased)
.RE
.RS 4
.I
ld\-lbfgs\-nocedal
\(hy None
.RE
.RS 4
.I
ld\-lbfgs
\(hy Low-storage BFGS
.RE
.RS 4
.I
ln\-praxis
\(hy Gradient-free Local Optimization via the Principal-Axis Method
.RE
.RS 4
.I
ld\-var1
\(hy Shifted Limited-Memory Variable-Metric, Rank 1
.RE
.RS 4
.I
ld\-var2
\(hy Shifted Limited-Memory Variable-Metric, Rank 2
.RE
.RS 4
.I
ld\-tnewton
\(hy Truncated Newton
.RE
.RS 4
.I
ld\-tnewton\-restart
\(hy Truncated Newton with steepest-descent restarting
.RE
.RS 4
.I
ld\-tnewton\-precond
\(hy Preconditioned Truncated Newton
.RE
.RS 4
.I
ld\-tnewton\-precond\-restart
\(hy Preconditioned Truncated Newton with steepest-descent restarting
.RE
.RS 4
.I
gn\-crs2\-lm
\(hy Controlled Random Search with Local Mutation
.RE
.RS 4
.I
ld\-mma
\(hy Method of Moving Asymptotes
.RE
.RS 4
.I
ln\-cobyla
\(hy Constrained Optimization BY Linear Approximation
.RE
.RS 4
.I
ln\-newuoa
\(hy Derivative-free Unconstrained Optimization by Iteratively Constructed Quadratic Approximation
.RE
.RS 4
.I
ln\-newuoa\-bound
\(hy Derivative-free Bound-constrained Optimization by Iteratively Constructed Quadratic Approximation
.RE
.RS 4
.I
ln\-neldermead
\(hy Nelder-Mead simplex algorithm
.RE
.RS 4
.I
ln\-sbplx
\(hy Subplex variant of Nelder-Mead
.RE
.RS 4
.I
ln\-bobyqa
\(hy Derivative-free Bound-constrained Optimization
.RE
.RS 4
.I
gn\-isres
\(hy Improved Stochastic Ranking Evolution Strategy
.RE
.RS 4
.I
auglag
\(hy Augmented Lagrangian algorithm
.RE
.RS 4
.I
auglag\-eq
\(hy Augmented Lagrangian algorithm with equality constraints only
.RE
.RS 4
.I
g\-mlsl
\(hy Multi-Level Single-Linkage (require local optimization and bounds)
.RE
.RS 4
.I
g\-mlsl\-lds
\(hy Multi-Level Single-Linkage (low-discrepancy-sequence, require local gradient based optimization and bounds)
.RE
.RS 4
.I
ld\-slsqp
\(hy Sequential Least-Squares Quadratic Programming
.RE

.RE
.RE
.RS 14
.I
step
= 0; double in [0, inf) 
.RS 2
Initial step size for gradient free methods.

.RE
.RE
.RS 14
.I
stop
= \-inf; double 
.RS 2
Stopping criterion: function value falls below this value.

.RE
.RE
.RS 14
.I
xtola
= 0; double in [0, inf) 
.RS 2
Stopping criterion: the absolute change of all x\-values is below  this value.

.RE
.RE
.RS 14
.I
xtolr
= 0; double in [0, inf) 
.RS 2
Stopping criterion: the relative change of all x\-values is below  this value.

.RE
.RE
.SH EXAMPLE
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'.
.HP
mia\-2dmyoica\-nonrigid2   \-i segment.set \-o registered.set \-k 2
.SH AUTHOR(s)
Gert Wollny
.SH COPYRIGHT
This software is Copyright (c) 1999\(hy2015 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'.