mia-3dprealign-nonrigid(1) Registration of a series of 3D images.

SYNOPSIS

mia-3dprealign-nonrigid -i <in-file> -o <out-file> [options]

DESCRIPTION

mia-3dprealign-nonrigid This program runs the non-rigid registration of an image series by first registering an already aligned subset of the images to one reference, and then by registering the remaining images by using synthetic references. The is a 3D version of G. Wollny, M-J Ledesma-Cabryo, P.Kellman, and A.Santos, "Exploiting Quasiperiodicity in Motion Correction of Free-Breathing," IEEE Transactions on Medical Imaging, 29(8), 2010.

OPTIONS

File-IO

-i --in-file=(input, required); io
input images following the naming pattern nameXXXX.ext For supported file types see PLUGINS:3dimage/io
-o --out-file=(output, required); io
file name base for registered files given as C-format string For supported file types see PLUGINS:3dimage/io
--save-references
Save synthetic references to files refXXXX.v

Preconditions & Preprocessing

-k --skip=0
Skip images at the begin of the seriesSkip images at the begin of the series
--preskip=20
Skip images at the beginning+skip of the series when searching for high contrats imageSkip images at the beginning+skip of the series when searching for high contrats image
--postskip=2
Skip images at the end of the series when searching for high contrats imageSkip images at the end of the series when searching for high contrats image
--max-candidates=20
maximum number of candidates for global reference imagemaximum number of candidates for global reference image
-S --cost-series=image:cost=[ngf:eval=ds]
Const function to use for the analysis of the seriesConst function to use for the analysis of the series For supported plugins see PLUGINS:3dimage/fullcost
--ref-idx=
save reference index number to this file
-R --global-reference=-1
save reference index number to this filesave reference index number to this file
-D --max-subset-delta=0
Maximum delta between two elements of the prealigned subsetMaximum delta between two elements of the prealigned subset

Registration

-O --optimizer=gsl:opt=gd,step=0.01
Optimizer used for minimizationOptimizer used for minimization For supported plugins see PLUGINS:minimizer/singlecost
-l --mr-levels=3
multi-resolution levelsmulti-resolution levels
-f --transForm=spline
transformation typetransformation type For supported plugins see PLUGINS:3dimage/transform
-1 --cost-subset=image:cost=[ngf:eval=ds]
Cost function for registration during the subset registrationCost function for registration during the subset registration For supported plugins see PLUGINS:3dimage/fullcost
-2 --cost-final=image:cost=ssd
Cost function for registration during the final registrationCost function for registration during the final registration For supported plugins see PLUGINS:3dimage/fullcost

Help & Info

-V --verbose=warning
verbosity of output, print messages of given level and higher priorities. Supported priorities starting at lowest level are:
info - Low level messages
trace - Function call trace
fail - Report test failures
warning - Warnings
error - Report errors
debug - Debug output
message - Normal messages
fatal - Report only fatal errors
--copyright
print copyright information
-h --help
print this help
-? --usage
print a short help
--version
print the version number and exit

Processing

--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).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).

PLUGINS: 1d/spacialkernel

cdiff
Central difference filter kernel, mirror boundary conditions are used.

(no parameters)
gauss
spacial Gauss filter kernel, supported parameters are:

w = 1; uint in [0, inf)
half filter width.

PLUGINS: 1d/splinebc

mirror
Spline interpolation boundary conditions that mirror on the boundary

(no parameters)
repeat
Spline interpolation boundary conditions that repeats the value at the boundary

(no parameters)
zero
Spline interpolation boundary conditions that assumes zero for values outside

(no parameters)

PLUGINS: 1d/splinekernel

bspline
B-spline kernel creation , supported parameters are:

d = 3; int in [0, 5]
Spline degree.

omoms
OMoms-spline kernel creation, supported parameters are:

d = 3; int in [3, 3]
Spline degree.

PLUGINS: 3dimage/combiner

absdiff
Image combiner 'absdiff'

(no parameters)
add
Image combiner 'add'

(no parameters)
div
Image combiner 'div'

(no parameters)
mul
Image combiner 'mul'

(no parameters)
sub
Image combiner 'sub'

(no parameters)

PLUGINS: 3dimage/cost

lncc
local normalized cross correlation with masking support., supported parameters are:

w = 5; uint in [1, 256]
half width of the window used for evaluating the localized cross correlation.

mi
Spline parzen based mutual information., supported parameters are:

cut = 0; float in [0, 40]
Percentage of pixels to cut at high and low intensities to remove outliers.

mbins = 64; uint in [1, 256]
Number of histogram bins used for the moving image.

mkernel = [bspline:d=3]; factory
Spline kernel for moving image parzen hinstogram. For supported plug-ins see PLUGINS:1d/splinekernel

rbins = 64; uint in [1, 256]
Number of histogram bins used for the reference image.

rkernel = [bspline:d=0]; factory
Spline kernel for reference image parzen hinstogram. For supported plug-ins see PLUGINS:1d/splinekernel

ncc
normalized cross correlation.

(no parameters)
ngf
This function evaluates the image similarity based on normalized gradient fields. Given normalized gradient fields $ _S$ of the src image and $ _R$ of the ref image various evaluators are implemented., supported parameters are:

eval = ds; dict
plugin subtype (sq, ds,dot,cross). Supported values are:
ds - square of scaled difference
dot - scalar product kernel
cross - cross product kernel

ssd
3D image cost: sum of squared differences, supported parameters are:

autothresh = 0; float in [0, 1000]
Use automatic masking of the moving image by only takeing intensity values into accound that are larger than the given threshold.

norm = 0; bool
Set whether the metric should be normalized by the number of image pixels.

ssd-automask
3D image cost: sum of squared differences, with automasking based on given thresholds, supported parameters are:

rthresh = 0; double
Threshold intensity value for reference image.

sthresh = 0; double
Threshold intensity value for source image.

PLUGINS: 3dimage/filter

bandpass
intensity bandpass filter, supported parameters are:

max = 3.40282e+38; float
maximum of the band.

min = 0; float
minimum of the band.

binarize
image binarize filter, supported parameters are:

max = 3.40282e+38; float
maximum of accepted range.

min = 0; float
minimum of accepted range.

close
morphological close, supported parameters are:

hint = black; string
a hint at the main image content (black|white).

shape = [sphere:r=2]; factory
structuring element. For supported plug-ins see PLUGINS:3dimage/shape

combiner
Combine two images with the given combiner operator. if 'reverse' is set to false, the first operator is the image passed through the filter pipeline, and the second image is loaded from the file given with the 'image' parameter the moment the filter is run., supported parameters are:

image =(input, required, string)
second image that is needed in the combiner.

op =(required, factory)
Image combiner to be applied to the images. For supported plug-ins see PLUGINS:3dimage/combiner

reverse = 0; bool
reverse the order in which the images passed to the combiner.

convert
image pixel format conversion filter, supported parameters are:

a = 1; float
linear conversion parameter a.

b = 0; float
linear conversion parameter b.

map = opt; dict
conversion mapping. Supported values are:
opt - apply a linear transformation that maps the real input range to the full output range
range - apply linear transformation that maps the input data type range to the output data type range
copy - copy data when converting
linear - apply linear transformation x -> a*x+b
optstat - apply a linear transform that maps based on input mean and variation to the full output range

repn = ubyte; dict
output pixel type. Supported values are:
none - no pixel type defined
float - floating point 32 bit
sbyte - signed 8 bit
ulong - unsigned 64 bit
double - floating point 64 bit
sint - signed 32 bit
ushort - unsigned 16 bit
sshort - signed 16 bit
uint - unsigned 32 bit
slong - signed 64 bit
bit - binary data
ubyte - unsigned 8 bit

crop
Crop a region of an image, the region is always clamped to the original image size in the sense that the given range is kept., supported parameters are:

end = [[4294967295,4294967295,4294967295]]; streamable
end of cropping range, maximum = (-1,-1,-1).

start = [[0,0,0]]; streamable
begin of cropping range.

dilate
3d image stack dilate filter, supported parameters are:

hint = black; string
a hint at the main image content (black|white).

shape = [sphere:r=2]; factory
structuring element. For supported plug-ins see PLUGINS:3dimage/shape

distance
Evaluate the 3D distance transform of an image. If the image is a binary mask, then result of the distance transform in each point corresponds to the Euclidian distance to the mask. If the input image is of a scalar pixel value, then the this scalar is interpreted as heighfield and the per pixel value adds to the distance.

(no parameters)
downscale
Downscale the input image by using a given block size to define the downscale factor. Prior to scaling the image is filtered by a smoothing filter to eliminate high frequency data and avoid aliasing artifacts., supported parameters are:

b = [[1,1,1]]; 3dbounds
blocksize.

bx = 1; uint in [1, inf)
blocksize in x direction.

by = 1; uint in [1, inf)
blocksize in y direction.

bz = 1; uint in [1, inf)
blocksize in z direction.

kernel = gauss; string
smoothing filter kernel to be applied, the size of the filter is estimated based on the blocksize..

erode
3d image stack erode filter, supported parameters are:

hint = black; string
a hint at the main image content (black|white).

shape = [sphere:r=2]; factory
structuring element. For supported plug-ins see PLUGINS:3dimage/shape

gauss
isotropic 3D gauss filter, supported parameters are:

w = 1; int in [0, inf)
filter width parameter.

gradnorm
3D image to gradient norm filter

(no parameters)
growmask
Use an input binary mask and a reference gray scale image to do region growing by adding the neighborhood pixels of an already added pixel if the have a lower intensity that is above the given threshold., supported parameters are:

min = 1; float
lower threshold for mask growing.

ref =(input, required, string)
reference image for mask region growing.

shape = 6n; factory
neighborhood mask. For supported plug-ins see PLUGINS:3dimage/shape

invert
intensity invert filter

(no parameters)
isovoxel
This filter scales an image to make the voxel size isometric and its size to correspond to the given value, supported parameters are:

interp = [bspline:d=3]; factory
interpolation kernel to be used . For supported plug-ins see PLUGINS:1d/splinekernel

size = 1; float in (0, inf)
isometric target voxel size.

kmeans
3D image k-means filter. In the output image the pixel value represents the class membership and the class centers are stored as attribute in the image., supported parameters are:

c = 3; int in [2, inf)
number of classes.

label
A filter to label the connected components of a binary image., supported parameters are:

n = 6n; factory
neighborhood mask. For supported plug-ins see PLUGINS:3dimage/shape

labelmap
Image filter to remap label id's. Only applicable to images with integer valued intensities/labels., supported parameters are:

map =(input, required, string)
Label mapping file.

labelscale
A filter that only creates output voxels that are already created in the input image. Scaling is done by using a voting algorithms that selects the target pixel value based on the highest pixel count of a certain label in the corresponding source region. If the region comprises two labels with the same count, the one with the lower number wins., supported parameters are:

out-size =(required, 3dbounds)
target size given as two coma separated values.

load
Load the input image from a file and use it to replace the current image in the pipeline., supported parameters are:

file =(input, required, string)
name of the input file to load from..

lvdownscale
This is a label voting downscale filter. It adownscales a 3D image by blocks. For each block the (non-zero) label that appears most times in the block is issued as output pixel in the target image. If two labels appear the same number of times, the one with the lower absolute value wins., supported parameters are:

b = [[1,1,1]]; 3dbounds
blocksize for the downscaling. Each block will be represented by one pixel in the target image..

mask
Mask an image, one image is taken from the parameters list and the other from the normal filter input. Both images must be of the same dimensions and one must be binary. The attributes of the image coming through the filter pipeline are preserved. The output pixel type corresponds to the input image that is not binary., supported parameters are:

input =(input, required, string)
second input image file name.

mean
3D image mean filter, supported parameters are:

w = 1; int in [1, inf)
half filter width.

median
median 3d filter, supported parameters are:

w = 1; int in [1, inf)
filter width parameter.

mlv
Mean of Least Variance 3D image filter, supported parameters are:

w = 1; int in [1, inf)
filter width parameter.

msnormalizer
3D image mean-sigma normalizing filter, supported parameters are:

w = 1; int in [1, inf)
half filter width.

open
morphological open, supported parameters are:

hint = black; string
a hint at the main image content (black|white).

shape = [sphere:r=2]; factory
structuring element. For supported plug-ins see PLUGINS:3dimage/shape

reorient
3D image reorientation filter, supported parameters are:

map = xyz; dict
oriantation mapping to be applied. Supported values are:
p-zxy - permutate x->y->z->x
r-x180 - rotate around x-axis clockwise 180 degree
xyz - keep orientation
p-yzx - permutate x->z->y->x
r-z180 - rotate around z-axis clockwise 180 degree
r-y270 - rotate around y-axis clockwise 270 degree
f-xz - flip x-z
f-yz - flip y-z
r-x90 - rotate around x-axis clockwise 90 degree
r-y90 - rotate around y-axis clockwise 90 degree
r-x270 - rotate around x-axis clockwise 270 degree
r-z270 - rotate around z-axis clockwise 270 degree
r-z90 - rotate around z-axis clockwise 90 degree
f-xy - flip x-y
r-y180 - rotate around y-axis clockwise 180 degree

resize
Resize an image. The original data is centered within the new sized image., supported parameters are:

size = [[0,0,0]]; streamable
new size of the image a size 0 indicates to keep the size for the corresponding dimension..

sandp
salt and pepper 3d filter, supported parameters are:

thresh = 100; float in [0, inf)
thresh value.

w = 1; int in [1, inf)
filter width parameter.

scale
3D image filter that scales to a given target size , supported parameters are:

interp = [bspline:d=3]; factory
interpolation kernel to be used . For supported plug-ins see PLUGINS:1d/splinekernel

s = [[0,0,0]]; 3dbounds
target size to set all components at once (component 0:use input image size).

sx = 0; uint in [0, inf)
target size in x direction (0:use input image size).

sy = 0; uint in [0, inf)
target size in y direction (0:use input image size).

sz = 0; uint in [0, inf)
target size in y direction (0:use input image size).

selectbig
A filter that creats a binary mask representing the intensity with the highest pixel count.The pixel value 0 will be ignored, and if two intensities have the same pixel count, then the result is undefined. The input pixel must have an integral pixel type.

(no parameters)
sepconv
3D image intensity separaple convolution filter, supported parameters are:

kx = [gauss:w=1]; factory
filter kernel in x-direction. For supported plug-ins see PLUGINS:1d/spacialkernel

ky = [gauss:w=1]; factory
filter kernel in y-direction. For supported plug-ins see PLUGINS:1d/spacialkernel

kz = [gauss:w=1]; factory
filter kernel in z-direction. For supported plug-ins see PLUGINS:1d/spacialkernel

sws
seeded watershead. The algorithm extracts exactly so many reagions as initial labels are given in the seed image., supported parameters are:

grad = 0; bool
Interpret the input image as gradient. .

mark = 0; bool
Mark the segmented watersheds with a special gray scale value.

n = [sphere:r=1]; factory
Neighborhood for watershead region growing. For supported plug-ins see PLUGINS:3dimage/shape

seed =(input, required, string)
seed input image containing the lables for the initial regions.

tee
Save the input image to a file and also pass it through to the next filter, supported parameters are:

file =(output, required, string)
name of the output file to save the image too..

thinning
3D morphological thinning, based on: Lee and Kashyap, 'Building Skeleton Models via 3-D Medial Surface/Axis Thinning Algorithms', Graphical Models and Image Processing, 56(6):462-478, 1994. This implementation only supports the 26 neighbourhood.

(no parameters)
transform
Transform the input image with the given transformation., supported parameters are:

file =(input, required, string)
Name of the file containing the transformation..

imgboundary = ; string
override image interpolation boundary conditions.

imgkernel = ; string
override image interpolator kernel.

variance
3D image variance filter, supported parameters are:

w = 1; int in [1, inf)
half filter width.

ws
basic watershead segmentation., supported parameters are:

evalgrad = 0; bool
Set to 1 if the input image does not represent a gradient norm image.

mark = 0; bool
Mark the segmented watersheds with a special gray scale value.

n = [sphere:r=1]; factory
Neighborhood for watershead region growing. For supported plug-ins see PLUGINS:3dimage/shape

thresh = 0; float in [0, 1)
Relative gradient norm threshold. The actual value threshold value is thresh * (max_grad - min_grad) + min_grad. Bassins separated by gradients with a lower norm will be joined.

PLUGINS: 3dimage/fullcost

image
Generalized image similarity cost function that also handles multi-resolution processing. The actual similarity measure is given es extra parameter., supported parameters are:

cost = ssd; factory
Cost function kernel. For supported plug-ins see PLUGINS:3dimage/cost

debug = 0; bool
Save intermediate resuts for debugging.

ref =(input, string)
Reference image.

src =(input, string)
Study image.

weight = 1; float
weight of cost function.

labelimage
Similarity cost function that maps labels of two images and handles label-preserving multi-resolution processing., supported parameters are:

maxlabel = 256; int in [2, 32000]
maximum number of labels to consider.

ref =(input, string)
Reference image.

src =(input, string)
Study image.

weight = 1; float
weight of cost function.

maskedimage
Generalized masked image similarity cost function that also handles multi-resolution processing. The provided masks should be densly filled regions in multi-resolution procesing because otherwise the mask information may get lost when downscaling the image. The mask may be pre-filtered - after pre-filtering the masks must be of bit-type.The reference mask and the transformed mask of the study image are combined by binary AND. The actual similarity measure is given es extra parameter., supported parameters are:

cost = ssd; factory
Cost function kernel. For supported plug-ins see PLUGINS:3dimage/maskedcost

ref =(input, string)
Reference image.

ref-mask =(input, string)
Reference image mask (binary).

ref-mask-filter = ; factory
Filter to prepare the reference mask image, the output must be a binary image.. For supported plug-ins see PLUGINS:3dimage/filter

src =(input, string)
Study image.

src-mask =(input, string)
Study image mask (binary).

src-mask-filter = ; factory
Filter to prepare the study mask image, the output must be a binary image.. For supported plug-ins see PLUGINS:3dimage/filter

weight = 1; float
weight of cost function.

taggedssd
Evaluates the Sum of Squared Differences similarity measure by using three tagged image pairs. The cost function value is evaluated based on all image pairs, but the gradient is composed by composing its component based on the tag direction., supported parameters are:

refx =(input, string)
Reference image X-tag.

refy =(input, string)
Reference image Y-tag.

refz =(input, string)
Reference image Z-tag.

srcx =(input, string)
Study image X-tag.

srcy =(input, string)
Study image Y-tag.

srcz =(input, string)
Study image Z-tag.

weight = 1; float
weight of cost function.

PLUGINS: 3dimage/io

analyze
Analyze 7.5 image

Recognized file extensions: .HDR, .hdr

 
Supported element types:
unsigned 8 bit, signed 16 bit, signed 32 bit, floating point 32 bit, floating point 64 bit

 
datapool
Virtual IO to and from the internal data pool

Recognized file extensions: .@

 
dicom
Dicom image series as 3D

Recognized file extensions: .DCM, .dcm

 
Supported element types:
signed 16 bit, unsigned 16 bit

 
hdf5
HDF5 3D image IO

Recognized file extensions: .H5, .h5

 
Supported element types:
binary data, signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, signed 64 bit, unsigned 64 bit, floating point 32 bit, floating point 64 bit

 
inria
INRIA image

Recognized file extensions: .INR, .inr

 
Supported element types:
signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, floating point 32 bit, floating point 64 bit

 
mhd
MetaIO 3D image IO using the VTK implementation (experimental).

Recognized file extensions: .MHA, .MHD, .mha, .mhd

 
Supported element types:
signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, floating point 32 bit, floating point 64 bit

 
nifti
NIFTI-1 3D image IO

Recognized file extensions: .NII, .nii

 
Supported element types:
signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, signed 64 bit, unsigned 64 bit, floating point 32 bit, floating point 64 bit

 
vff
VFF Sun raster format

Recognized file extensions: .VFF, .vff

 
Supported element types:
unsigned 8 bit, signed 16 bit

 
vista
Vista 3D

Recognized file extensions: .V, .VISTA, .v, .vista

 
Supported element types:
binary data, signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, floating point 32 bit, floating point 64 bit

 
vti
3D image VTK-XML in- and output (experimental).

Recognized file extensions: .VTI, .vti

 
Supported element types:
signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, floating point 32 bit, floating point 64 bit

 
vtk
3D VTK image legacy in- and output (experimental).

Recognized file extensions: .VTK, .VTKIMAGE, .vtk, .vtkimage

 
Supported element types:
binary data, signed 8 bit, unsigned 8 bit, signed 16 bit, unsigned 16 bit, signed 32 bit, unsigned 32 bit, floating point 32 bit, floating point 64 bit

 

PLUGINS: 3dimage/maskedcost

lncc
local normalized cross correlation with masking support., supported parameters are:

w = 5; uint in [1, 256]
half width of the window used for evaluating the localized cross correlation.

mi
Spline parzen based mutual information with masking., supported parameters are:

cut = 0; float in [0, 40]
Percentage of pixels to cut at high and low intensities to remove outliers.

mbins = 64; uint in [1, 256]
Number of histogram bins used for the moving image.

mkernel = [bspline:d=3]; factory
Spline kernel for moving image parzen hinstogram. For supported plug-ins see PLUGINS:1d/splinekernel

rbins = 64; uint in [1, 256]
Number of histogram bins used for the reference image.

rkernel = [bspline:d=0]; factory
Spline kernel for reference image parzen hinstogram. For supported plug-ins see PLUGINS:1d/splinekernel

ncc
normalized cross correlation with masking support.

(no parameters)
ssd
Sum of squared differences with masking.

(no parameters)

PLUGINS: 3dimage/shape

18n
18n neighborhood 3D shape creator

(no parameters)
26n
26n neighborhood 3D shape creator

(no parameters)
6n
6n neighborhood 3D shape creator

(no parameters)
sphere
Closed spherical shape neighborhood including the pixels within a given radius r., supported parameters are:

r = 2; float in (0, inf)
sphere radius.

PLUGINS: 3dimage/transform

affine
Affine transformation (12 degrees of freedom), supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

axisrot
Restricted rotation transformation (1 degrees of freedom). The transformation is restricted to the rotation around the given axis about the given rotation center, supported parameters are:

axis =(required, 3dfvector)
rotation axis.

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

origin =(required, 3dfvector)
center of the transformation.

raffine
Restricted affine transformation (3 degrees of freedom). The transformation is restricted to the rotation around the given axis and shearing along the two axis perpendicular to the given one, supported parameters are:

axis =(required, 3dfvector)
rotation axis.

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

origin =(required, 3dfvector)
center of the transformation.

rigid
Rigid transformation, i.e. rotation and translation (six degrees of freedom)., supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

origin = [[0,0,0]]; 3dfvector
Relative rotation center, i.e. <0.5,0.5,0.5> corresponds to the center of the volume.

rotation
Rotation transformation (three degrees of freedom)., supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

origin = [[0,0,0]]; 3dfvector
Relative rotation center, i.e. <0.5,0.5,0.5> corresponds to the center of the volume.

rotbend
Restricted transformation (4 degrees of freedom). The transformation is restricted to the rotation around the x and y axis and a bending along the x axis, independedn in each direction, with the bending increasing with the squared distance from the rotation axis., supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

norot = 0; bool
Don't optimize the rotation.

origin =(required, 3dfvector)
center of the transformation.

spline
Free-form transformation that can be described by a set of B-spline coefficients and an underlying B-spline kernel., supported parameters are:

anisorate = [[0,0,0]]; 3dfvector
anisotropic coefficient rate in pixels, nonpositive values will be overwritten by the 'rate' value..

debug = 0; bool
enable additional debuging output.

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

kernel = [bspline:d=3]; factory
transformation spline kernel. For supported plug-ins see PLUGINS:1d/splinekernel

penalty = ; factory
transformation penalty energy term. For supported plug-ins see PLUGINS:3dtransform/splinepenalty

rate = 10; float in [1, inf)
isotropic coefficient rate in pixels.

translate
Translation (three degrees of freedom), supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

vf
This plug-in implements a transformation that defines a translation for each point of the grid defining the domain of the transformation., supported parameters are:

imgboundary = mirror; factory
image interpolation boundary conditions. For supported plug-ins see PLUGINS:1d/splinebc

imgkernel = [bspline:d=3]; factory
image interpolator kernel. For supported plug-ins see PLUGINS:1d/splinekernel

PLUGINS: 3dtransform/io

bbs
Binary (non-portable) serialized IO of 3D transformations

Recognized file extensions: .bbs

 
datapool
Virtual IO to and from the internal data pool

Recognized file extensions: .@

 
vista
Vista storage of 3D transformations

Recognized file extensions: .v, .v3dt

 
xml
XML serialized IO of 3D transformations

Recognized file extensions: .x3dt

 

PLUGINS: 3dtransform/splinepenalty

divcurl
divcurl penalty on the transformation, supported parameters are:

curl = 1; float in [0, inf)
penalty weight on curl.

div = 1; float in [0, inf)
penalty weight on divergence.

norm = 0; bool
Set to 1 if the penalty should be normalized with respect to the image size.

weight = 1; float in (0, inf)
weight of penalty energy.

PLUGINS: minimizer/singlecost

gdas
Gradient descent with automatic step size correction., supported parameters are:

ftolr = 0; double in [0, inf)
Stop if the relative change of the criterion is below..

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

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

min-step = 0.1; double in (0, inf)
Minimal absolute step size.

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

gdsq
Gradient descent with quadratic step estimation, supported parameters are:

ftolr = 0; double in [0, inf)
Stop if the relative change of the criterion is below..

gtola = 0; double in [0, inf)
Stop if the inf-norm of the gradient is below this value..

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

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

step = 0.1; double in (0, inf)
Initial step size.

xtola = 0; double in [0, inf)
Stop if the inf-norm of x-update is below this value..

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 in (0, inf)
gradient based optimizers: stop when |grad| < eps, simplex: stop when simplex size < eps..

iter = 100; uint in [1, inf)
maximum number of iterations.

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

step = 0.001; double in (0, inf)
initial step size.

tol = 0.1; double in (0, inf)
some tolerance parameter.

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 in [0, inf)
Stopping criterion: the absolute change of the objective value is below this value.

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

higher = inf; double
Higher boundary (equal for all parameters).

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

lower = -inf; double
Lower boundary (equal for all parameters).

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

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

step = 0; double in [0, inf)
Initial step size for gradient free methods.

stop = -inf; double
Stopping criterion: function value falls below this value.

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

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

EXAMPLE

Register the image series given by images imageXXXX.v by optimizing a spline based transformation with a coefficient rate of 16 pixel, skipping two images at the beginning and using normalized gradient fields as initial cost measure and SSD as final measure. Penalize the transformation by using divcurl with aweight of 2.0. As optimizer an nlopt based newton method is used.
mia-3dprealign-nonrigid mia-3dprealign-nonrigid -i imageXXXX.v -o registered -t vista -k 2-F spline:rate=16,penalty=[divcurl:weight=2] -1 image:cost=[ngf:eval=ds] -2 image:cost=ssd -O nlopt:opt=ld-var1,xtola=0.001,ftolr=0.001,maxiter=300

AUTHOR(s)

Gert Wollny

COPYRIGHT

This software is Copyright (c) 1999-2015 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'.