Sisyphe.core.sisypheImage

External packages/modules

Cython, static compiler, https://cython.org/

ANTs, image registration, http://stnava.github.io/ANTs/

DIPY, MR diffusion image processing, https://www.dipy.org/

ITK, medical image processing, https://itk.org/

NiBabel, nibabel image class access, https://nipy.org/nibabel

Numpy, scientific computing, https://numpy.org/

PyQt5, Qt GUI, https://www.riverbankcomputing.com/software/pyqt/

scikit-image, image processing, https://scikit-image.org/

Scipy, scientific computing, https://docs.scipy.org

SimpleITK, medical image processing, https://simpleitk.org/

vtk, visualization engine/3D rendering, https://vtk.org/

Description

Specialized SisypheImage class with unsigned short datatype.

Inheritance

object -> SisypheImage -> SisypheBinaryImage

Creation: 12/01/2021 Last revision: 22/06/2024

copyFromANTSImage(img: ANTsImage) → None

Copy ANTs image buffer to current instance.

Parameters

imgants.core.ANTsImage: image to copy

copyFromITKImage(img: <itkTemplate itk::Image>) → None

Copy ITK image buffer to current instance.

Parameters

imgitk.Image: image to copy

copyFromNumpyArray(img: ndarray, spacing: list[float] | tuple[float, float, float] = (1.0, 1.0, 1.0), origin: list[float] | tuple[float, float, float] = (0.0, 0.0, 0.0), direction: tuple | list = (1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0), defaultshape: bool = True)

Copy Numpy array buffer to current instance.

Parameters

imgnumpy.ndarray: image to copy
spacingtuple[float, float, float] | list[float, float, float]: voxel sizes in mm (default 1.0, 1.0, 1.0)
origintuple[float, float, float] | list[float, float, float]: origin coordinates (default 0.0, 0.0, 0.0)
directiontuple[float, …] | list[float]: axes directions
defaultshapebool: if True, numpy shape order is reversed (i.e. z, y, x)

copyFromSITKImage(img: Image) → None

Copy SimpleITK image buffer to current instance.

Parameters

imgSimpleITK.Image: image to copy

copyFromVTKImage(img: vtkImageData) → None

Copy VTK image buffer to current instance.

Parameters

imgvtk.vtkImageData: image to copy

setSITKImage(img: Image) → None

Shallow copy of a SimpleITK Image to the SimpleITK Image attribute of the current SisypheImage instance.

Parameters

imgSimpleITK.Image: image to copy

Description

PySisyphe image class.

This class provides access to internal SimpleITK, ITK, VTK and numpy image classes, which share the same image buffer.

Overloaded operators which work with int, float, SisypheVolume, SisypheImage, SimpleITK Image, and numpy array:

Arithmetic +, -, /, //, *

Logic & (and), | (or), ^ (xor), ~ (not)

Relational >, >=, <, <=, ==, !=

Getter and Setter access to scalar values with slicing ability. Getter: v = instance_name[x, y, z] Setter: instance_name[x, y, z] = v x, y, z are int or slice object (syntax first:last:step)

Scope of methods:

copy to/from SimpleITK, ITK, VTK, ANTs, NiBabel and numpy image instances

various image attributes getter methods

descriptive statistics (minimum, maximum, median, mean, standard-deviation, percentile…)

IO methods for common neuroimaging file formats (BrainVoyager, Nifti, Nrrd, Minc, VTK)

Inheritance

object -> SisypheImage

Creation: 12/01/2021 Last revision: 08/04/2025

allocate(matrix: list[int] | tuple[int, int, int], datatype: str) → None

Initialize image buffer.

Parameters

matrixlist[float, float, float] | tuple[float, float, float]: image size in x, y, z dimensions
datatypestr: numpy datatype

cast(datatype: str) → tuple[SisypheImage, float, float]

SisypheImage copy of the current SisypheImage instance with a new datatype.

Parameters

datatypestr: numpy datatype

Returns

tuple[SisypheImage, float, float]

first float, slope
second float, intercept
linear transform: cast value = slope * original value + intercept, linear transform is applied if direct cast is not possible (overflow error)

copy() → SisypheImage

SisypheImage copy of the current SisypheImage instance.

Returns

SisypheImage: image copy

copyComponent(c: int = 0) → SisypheImage

Extract single-component SisypheImage from the current multi-component SisypheImage instance.

Parameters

cint: component index (default 0)

Returns

SisypheImage: extracted single-component image

copyFromANTSImage(img: ANTsImage) → None

Copy an ANTsImage buffer to the current SisypheImage instance. Image buffer is not shared between ANTsImage image and SisypheImage instances.

Parameters

imgants.core.ANTsImage: image to copy

copyFromITKImage(img: <itkTemplate itk::Image>) → None

Copy an ITKImage buffer to the current SisypheImage instance. Image buffer is not shared between ITKImage image and SisypheImage instances.

Parameters

imgitk.Image: image to copy

copyFromNibabelImage(img: Nifti1Image): Copy a Nifti1Image to the current SisypheImage instance. Image buffer is not shared between Nifti1Image and SisypheImage instances (deep copy).

Parameters

img : nibabel.nifti1.Nifti1Image

copyFromNumpyArray(img: ndarray, spacing: list[float] | tuple[float, float, float] = (1.0, 1.0, 1.0), origin: list[float] | tuple[float, float, float] = (0.0, 0.0, 0.0), direction: tuple | list = (1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0), defaultshape: bool = True) → None

Copy a Numpy array buffer to the current SisypheImage instance. Image buffer is not shared between numpy array and SisypheImage instances.

Parameters

imgnumpy.ndarray

image to copy

spacinglist[float, float, float] | tuple[float, float, float],

voxel sizes in mm (default 1.0, 1.0, 1.0)

originlist[float, float, float] | tuple[float, float, float],

origin coordinates (default 0.0, 0.0, 0.0)

directionlist[float]

axes directions

defaultshapebool

3D: if True, numpy array (z, y, x) shape, otherwise (x, y, z) shape
4D: if True, numpy array (n, z, y, x) shape, otherwise (x, y, z, n) shape

copyFromSITKImage(img: Image) → None

Copy a SimpleITK image buffer to the current SisypheImage instance. Image buffer is not shared between SimpleITK image and SisypheImage instances.

Parameters

imgSimpleITK.Image: image to copy

copyFromVTKImage(img: vtkImageData) → None

Copy a VTKImageData buffer to the current SisypheImage instance. Image buffer is not shared between VTKImageData image and SisypheImage instances.

Parameters

imgvtk.vtkImageData: image to copy

copyToANTSImage(dtype: str = '', direction: bool = False) → ANTsImage

ANTsImage copy of the current SisypheImage instance. Image buffer is not shared between ANTsImage and SisypheImage instances.

Parameters

dtypestr: datatype conversion (default no conversion)
directionbool: set direction to LPI+ if True (default False)

Returns

ants.core.ANTSImage: image copy

copyToITKImage() → <itkTemplate itk::Image>

ITKImage copy of the current SisypheImage instance. Image buffer is not shared between ITKImage and SisypheImage instances.

Returns

itk.Image: image copy

copyToNibabelImage() → Nifti1Image

Nibabel Nifti1Image copy of the current SisypheImage instance. Image buffer is not shared between Nifti1Image and SisypheImage instances (deep copy).

Returns

nibabel.nifti1.Nifti1Image: image copy

copyToNumpyArray(defaultshape: bool = True) → ndarray

Numpy array copy of the current SisypheImage instance. Image buffer is not shared between numpy array and SisypheImage instances.

Parameters

defaultshapebool

3D: if True returns (z, y, x) shape, otherwise returns shape (x, y, z)
4D: if True returns (n, z, y, x) shape, otherwise returns shape (x, y, z, n)

Returns

numpy.ndarray: image copy

copyToSITKImage() → Image

SimpleITK image copy of the current SisypheImage instance. Image buffer is not shared between SimpleITK image and SisypheImage instances.

Returns

SimpleITK.Image: image copy

copyToVTKImage() → vtkImageData

VTKImageData copy of the current SisypheImage instance. Image buffer is not shared between VTKImageData and SisypheImage instances.

Returns

vtk.vtkImageData: image copy

fillWith(v: float = 0.0) → None

Fill the entire image with a given value.

Parameters

vfloat: value to fill the image with (default 0.0)

flip(axis: list[bool]) → None

Flip the current SisypheImage instance.

Parameters

axislist[bool]: flip axis if True, axis order x, y, z.

getArgmax(mask: str | SisypheImage | SisypheBinaryImage | None = None) → tuple[int, ...] | list[tuple[int, ...]]

Get the coordinates of the highest scalar value.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

Returns

tuple[int, int, int] | list[tuple[int, int, int]]: x, y, z or list[(x, y, z)] for multicomponent image

getBackgroundThreshold(algo: str = 'mean', c: int | None = 0) → float

Calc threshold to segment background.

Parameters

algostr

Automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

float: threshold to segment background

getBoundingBox(threshold: float = 0.0, margin: int = 0, blobs: bool = False) → ndarray

Get the bounding box a thresholded image of the current SisypheVolume instance.

Parameters

thresholdfloat: threshold applied to scalar image (default 0.0)
marginint: isotropic margin added to the bounding box (in voxels, default 0 i.e. non margin)
blobsbool: bounding box of the whole image if False, of each image blob if True

Returns

ndarray

if blobs is False, 6 elements ndarray: xmin, xmax, ymin, ymax, zmin, zmax
if blobs is True, 6 elements ndarray for each blob of the thresholded image

getCenter() → list[float] | tuple[float, float, float]

Get center of the volume as world coordinates.

Returns

tuple[float, float, float]: center of the volume

getCentroid(binary: bool = False, c: int | None = 0) → tuple[float, ...]

Get centroid of the image.

Parameters

binarybool

Processing on a binary mask if True, or greyscale image (as weights) if False

cint | None

parameter only used for multi-component image
int, index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

tuple[float, …]: centroid coordinates

getComponentAllZero(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get vectors whose values are all zero in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentArgmax() → SisypheImage

Get the component index of the highest scalar value of each vector in a multicomponent image.

Returns

SisypheImage: single-component image

getComponentArgmin(nonzero: bool = False) → SisypheImage

Get the component index of the lowest scalar value of each vector in a multicomponent image.

Parameters

nonzerobool: exclude zero values if True

Returns

SisypheImage: single-component image

getComponentKurtosis(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get kurtosis of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentMax(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get maximum of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentMean(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get mean of each vector in a multicomponent image. Process all components or a subset selected by indices.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentMedian(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get median of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentMin(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get minimum of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentNumberOfNonZero(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get the number of non-zero values for each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentPercentile(perc: int = 25, c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get percentile of each vector in a multicomponent image.

Parameters

percint

percentile (default 25, first quartile)

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentRange(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get range (max - min) of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentSkewness(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get skewness of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getComponentStd(c: list[int] | tuple[int, ...] | slice | None = None) → SisypheImage

Get standard deviation of each vector in a multicomponent image.

Parameters

cint | list[int] | tuple[int, …] | slice | None

component indices

list[int]: list of component indices
tuple[int, …]: tuple of component indices
slice: slice of component indices (start:end:step)
None: no selection, all components are processed (default)

Returns

SisypheImage: single-component image

getCrop(c: tuple | list | ndarray) → SisypheImage

Get a cropped image of the current SisypheVolume instance.

Parameters

ctuple | list: xmin, xmax, ymin, ymax, zmin, zmax

Returns

SisypheImage: cropped image

getCroppedProjection(slc: int, d: str = 'left', thickness: float = 0.0, func: str = 'max', output: str = 'native', emask: SisypheImage | None = None) → Image | ndarray | SisypheImage

Get a 2D projection of the current SisypheVolume instance. The projection is processed to a depth expressed in mm from the head surface (default 0.0, whole brain, no thickness). Operators applied to voxels on a projection line: maximum, mean, median, standard deviation, sum. The volume can be cut out to a specified depth (slice index) in the direction of projection.

Parameters

slcint: cutting plan (orientation is given by d parameter, see below)
dstr: direction of the projection: ‘left’, ‘right’, ‘ant’, ‘post’, ‘top’, ‘bottom’
thickness: float: projection by a given thickness in mm (default 0.0, whole brain, no thickness)
funcstr: ‘max’, ‘mean’, ‘median’, ‘std’, ‘sum’
outputstr: output format: ‘numpy’ numpy.ndarray, ‘sitk’ SimpleITK.Image, ‘native’ SisypheImage
emaskSisypheImage: if mask is None (default), mask is processed from the current SisypheImage instance (automatic thresholding with Huang algorithm for grayscale images, or non-zero voxels for label images). if an explicit mask is given, array size must match the current SisypheImage instance.

Returns

SimpleITK.Image | numpy.ndarray | SisypheImage

SimpleITK.Image if func = ‘sitk’,
numpy.ndarray if func = ‘numpy’,
SisypheImage if func = ‘native’

getDatatype() → str

Get image datatype as numpy datatype (i.e. ‘uint8’, ‘int8’, ‘uint16’, ‘int16’, ‘int32’, ‘uint32’, ‘int64’, ‘uint64’, ‘float32’, ‘float62’).

Returns

str: numpy datatype

getDatatypeAs(lib: str = 'sitk') → str | int | None

Get datatype with the library format of ANTs, SimpleITK, ITK or VTK.

Parameters

libstr | int

library name (i.e. ‘ants’, ‘itk’, ‘sitk’, ‘vtk’)
library code (i.e. ants=0, itk=1, sitk=2, vtk=3)

Returns

str | int: datatype (str or int code)

getDepth() → int

Get image size in z.

Returns

int: z-axis image size

getDirections() → tuple[float, ...]

Get vectors of image axes in RAS+ coordinates system.

PySisyphe uses RAS+ world coordinates system convention (as MNI, Nibabel, Dipy…) with origin to corner of the voxel:

x, direction [1.0, 0.0, 0.0]: left(-) to right(+)

y, direction [0.0, 1.0, 0.0]: posterior(-) to anterior(+)

z, direction [0.0, 0.0, 1.0]: inferior(-) to superior(+)

Directions is a list of 9 float, 3 vectors of 3 floats:

First vector, x-axis image direction, [1.0, 0.0, 0.0] (RAS+ default)

Second vector, y-axis image direction, [0.0, 1.0, 0.0] (RAS+ default)

Third vector, z-axis image direction, [0.0, 0.0, 1.0] (RAS+ default)

Returns

tuple[float, …]: vectors of image axes

getFieldOfView(decimals: int | None = None) → tuple[float, float, float]

Get field of view of the volume.

Parameters

decimalsint | None: number of decimals used to round values, if None no round

Returns

tuple[float, float, float]: field of view

getFirstVectorDirection() → list[float] | tuple[float, float, float]

Get first direction vector, x image axis in RAS+ coordinates system.

PySisyphe uses RAS+ world coordinates system convention (as MNI, Nibabel, Dipy…) with origin to corner of the voxel:

x, direction [1.0, 0.0, 0.0]: left(-) to right(+)

y, direction [0.0, 1.0, 0.0]: posterior(-) to anterior(+)

z, direction [0.0, 0.0, 1.0]: inferior(-) to superior(+)

Directions is a list of 9 float, 3 vectors of 3 floats:

First vector, x-axis image direction, [1.0, 0.0, 0.0] (RAS+ default)

Second vector, y-axis image direction, [0.0, 1.0, 0.0] (RAS+ default)

Third vector, z-axis image direction, [0.0, 0.0, 1.0] (RAS+ default)

Returns

tuple[float, float, float]: vector of the x-axis image

getFlip(axis: list[bool]) → SisypheImage

Get a flipped copy of the current SisypheImage instance.

Parameters

axislist[bool]: flip axis if True, axis order x, y, z.

Returns

SisypheImage: flipped image

getHeight() → int

Get image size in y.

Returns

int: y-axis image size

getHistogram(mask: str | SisypheImage | SisypheBinaryImage | None = None, bins: int = 100, range: tuple[float, float] | None = None, density: bool = False, c: int = 0) → ndarray

Get histogram of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

binsint

defines the number of equal-width bins in the given range.

rangetuple[float, float] | None

lower and upper range of the bins. If not provided (default None), range is simply (a.min(), a.max()), values outside the range are ignored.

densitybool

if False (default), the result will contain the number of samples in each bin. If True, the result is the value of the probability density function at the bin, normalized such that the integral over the range is 1.

cint

component index

Returns

ndarray: histogram

getITKImage()

ITKImage view (pointer) of the current SisypheImage. Image buffer is shared between ITKImage and current SisypheImage instances.

Returns

itk.Image: shallow copy of image

Get kurtosis of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int, index of the component to process
None, all components processed. In this case, the method returns a list of kurtosis values of each

component of the multi-component image

Returns

float | list[float]

kurtosis value for single-component image
kurtosis value for multi-component image, if component index c is int (default 0)
list of kurtosis values of each component of a multi-component image, if component index c is None

getLocalMaxima(threshold: float) → ndarray

Get the local maxima of the current SisypheVolume instance.

Determine all maxima of the image with height >= threshold. The local maxima are defined as connected sets of voxels with equal scalar value strictly greater than the scalar value of all voxels in direct neighborhood of the set. A local maximum M of height threshold is a local maximum for which there is at least one path joining M with an equal or higher local maximum on which the minimal value is f(M) - h (i.e. the values along the path are not decreasing by more than h with respect to the maximum’s value) and no path to an equal or higher local maximum for which the minimal value is greater.

Parameters

thresholdfloat: local maxima threshold

Returns

ndarray: coordinates of the local maxima

getMask(algo: str = 'huang', morpho: str = '', niter: int = 1, kernel: int = 0, fill: str = '2d', c: int | None = 0) → SisypheImage

Calc SisypheImage mask of the head.

Processing stages : 1. automatic thresholding of background 2. binary not of background = object 2. optional stage, iterative binary morphology 3. keep major blob 4. fill holes

Parameters

algostr

Automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’

morphostr

binary morphology operator, ‘dilate’, ‘erode’, ‘open’, ‘close’ or ‘’ (default, no morphology)

niterint

number of binary morphology iterations (default 1)

kernelint

structuring element size (default 0, no morphology)

fillstr

‘2d’, fill holes slice by slice
‘3d’, fill holes in 3D
‘’, no filling (default)

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

SisypheImage: mask

getMask2(algo: str = 'huang', morphoiter: int = 2, kernel: int = 0, c: int | None = 0) → SisypheImage

Calc SisypheImage mask of the head.

Processing stages : 1. automatic thresholding of background 2. erode background 3. keep major blob (i.e.remove blobs in head/object) 4. dilate background 5. binary not of background = object 6. erode object 7. keep major blob (i.e.remove background blobs in head/object) 8. dilate object

Parameters

algostr

Automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’

morphoiterint

number of binary morphology iterations (default 2)

kernelint

structuring element size, 0 automatic value (kernel=2 if spacing < 1.5 mm, kernel=1 otherwise)

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

SisypheImage: mask

getMaskROI(algo: str = 'huang', morpho: str = '', niter: int = 1, kernel: int = 0, fill: str = '2d', c: int | None = 0) → SisypheROI

Calc Sisyphe.core.sisypheROI.SisypheROI mask of the head.

Parameters

algostr

Automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’

morphostr

binary morphology operator, ‘dilate’, ‘erode’, ‘open’, ‘close’ or ‘’ (default, no morphology)

niterint

number of binary morphology iterations (default 1)

kernelint

structuring element size (default 0, no morphology)

fillstr

‘2d’, fill holes slice by slice
‘3d’, fill holes in 3D
‘’, no filling

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

Sisyphe.core.sisypheROI.SisypheROI: mask

getMaskROI2(algo: str = 'huang', morphoiter: int = 2, kernel: int = 0, c: int | None = 0) → SisypheROI

Calc Sisyphe.core.sisypheROI.SisypheROI mask of the head.

Parameters

algostr

Automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’

morphoiterint

number of binary morphology iterations

kernelint

structuring element size, 0 automatic value (kernel=2 if spacing < 1.5 mm, kernel=1 otherwise)

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

Sisyphe.core.sisypheROI.SisypheROI: mask

Get max of image. Calculation can only be performed on non-zero values.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int, index of the component to process
None, all components processed. In this case, the method returns a list of max values of each component

of the multi-component image

Returns

float | list[float]

max value for single-component image
max value for multi-component image, if component index c is int (default 0)
list of max values of each component of a multi-component image, if component index c is None

Get mean of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int index of the component to process
None, all components processed. In this case, the method returns a list of mean values of each component

of the multi-component image

Returns

float | list[float]

mean value for single-component image
mean value for multi-component image, if component index c is int (default 0)
list of mean values of each component of a multi-component image, if component index c is None

Get median of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int index of the component to process
None, all components processed. In this case, the method returns a list of median values of each component

of the multi-component image

Returns

float | list[float]

median value for single-component image
median value for multi-component image, if component index c is int (default 0)
list of median values of each component of a multi-component image, if component index c is None

getMemorySize(rep: str = 'B') → int | float

Get the memory size of the current SisypheImage instance.

Parameters

repstr: memory unit: ‘B’ Bytes, ‘KB’ Kilobytes or ‘MB’ Megabytes

Returns

int | float,

int if rep == ‘B’
float if rep in (‘KB’, ‘MB’)

getMin(nonzero: bool = False, c: int | None = 0) → float | list[float]

Get min of image. Processing can be performed on non-zero values.

Parameters

nonzerobool

exclude zero values if True

cint | None

parameter only used for multi-component image
int index of the component to process
None, all components processed. In this case, the method returns a list of min values of each component

of the multi-component image

Returns

float | list[float]

min value for single-component image
min value for multi-component image, if component index c is int (default 0)
list of min values of each component of a multi-component image, if component index c is None

getNative2DOrientation() → int

Get code of the native orientation (0 3D, 1 axial, 2 coronal, 3 sagittal).

Returns

int: code of the native orientation

getNative2DOrientationAsString() → tuple[str, str | None]

Get the native orientation as str.

Returns

tuple[str, str]

first str, mode ‘2D’ or ‘3D’
second str, orientation ‘Axial’, ‘Coronal’ or ‘Sagittal’

getNonZeroMask(c: int | None = 0) → SisypheImage

Calc SisypheImage mask of non-zero voxels of the current SisypheImage instance.

Parameters

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

SisypheImage: mask

getNumberOfComponentsPerPixel() → int

Get number of components. Array element of single component image is a scalar, array element of multi-component image is a vector. The number of components is the vector element count.

Returns

int: number of components

getNumberOfDimensions() → int

Get number of dimensions.

Returns

int: number of dimensions

getNumberOfNonZero(c: int = 0) → int

Get number of non-zero in image.

Parameters

cint: component index

Returns

int: number of non-zero scalar values

getNumberOfVoxels() → int

Get voxel count in the array.

Returns

int: voxel count

getNumberOfVoxelsInXYPlane() → int

Get voxel count in axial slice.

Returns

int: voxel count in XY plane

getNumberOfVoxelsInXZPlane() → int

Get voxel count in coronal slice.

Returns

int: voxel count in XZ plane

getNumberOfVoxelsInYZPlane() → int

Get voxel count in sagittal slice.

Returns

int: voxel count in YZ plane

getNumpy(defaultshape: bool = True) → ndarray

Numpy array view (pointer) of the current SisypheImage instance. Image buffer is shared between numpy array and current SisypheImage instances.

Parameters

defaultshapebool

3D: if True returns (z, y, x) shape, otherwise returns shape (x, y, z)
4D: if True returns (n, z, y, x) shape, otherwise returns shape (x, y, z, n)

Returns

numpy.ndarray: shallow copy of image

getOrigin() → list[float] | tuple[float, float, float]

Get geometrical reference origin coordinates.

Returns

tuple[float, float, float]: image origin in world coordinates

Get percentile of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

percint

percentile (default 25, first quartile)

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int, index of the component to process
None, all components processed. In this case, the method returns a list of percentile values of each

component of the multi-component image

Returns

float | list[float]

percentile value for single-component image
percentile value for multi-component image, if component index c is int (default 0)
list of percentile values of each component of a multi-component image, if component index c is None

getProjection(d: str = 'left', thickness: float = 0.0, func: str = 'max', output: str = 'native', emask: SisypheImage | None = None) → Image | ndarray | SisypheImage

Get a 2D projection of the current SisypheImage instance. The projection is processed to a depth expressed in mm from the head surface (default 0.0, whole brain, no thickness). Operators applied to voxels on a projection line: maximum, mean, median, standard deviation, sum.

Parameters

dstr: direction of the projection: ‘left’, ‘right’, ‘ant’, ‘post’, ‘top’, ‘bottom’
thickness: float: projection by a given thickness in mm (default 0.0, whole brain, no thickness)
funcstr: ‘max’, ‘mean’, ‘median’, ‘std’, ‘sum’, ‘label’
outputstr: output format: ‘numpy’ numpy.ndarray, ‘sitk’ SimpleITK.Image, ‘native’ SisypheImage
emaskSisypheImage: if mask is None (default), mask is processed from the current SisypheImage instance (automatic thresholding with Huang algorithm for grayscale images, or non-zero voxels for label images). if an explicit mask is given, array size must match the current SisypheImage instance.

Returns

SimpleITK.Image | numpy.ndarray | SisypheImage

SimpleITK.Image if func = ‘sitk’,
numpy.ndarray if func = ‘numpy’,
SisypheImage if func = ‘native’

getROI(threshold: float = 0.5, op: str = '>', c: int | None = 0) → SisypheROI

Converting current SisypheImage instance to Sisyphe.core.sisypheROI.SisypheROI using a threshold.

Parameters

thresholdfloat

threshold for binarization (default 0.5)

opstr

comparison operator: ‘>’ (default), ‘>=’, ‘<’, ‘<=’, ‘==’, ‘!=’, ‘0<’ (0 < x < threshold), ‘0<=’ (0 < x <= threshold)

cint | None

parameter only used for multi-component image
int index of the component to process (default 0, first component)
None, processing is performed on the mean image

Returns

Sisyphe.core.sisypheROI.SisypheROI

Get range (min, max) of image.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int, index of the component to process
None, all components processed. In this case, the method returns a list of range values of each component

of the multi-component image

Returns

tuple[float, float] | list[tuple[float, float]]

(min, max) values for single-component image
(min, max) values for multi-component image, if component index c is int (default 0)
list of max values of each component of a multi-component image, if component index c is None

getSITKImage() → Image

SimpleITK view (pointer) of the current SisypheImage instance. Image buffer is shared between SimpleITK image and current SisypheImage instances.

Returns

SimpleITK.Image: shallow copy of image

getSecondVectorDirection() → list[float] | tuple[float, float, float]

Get second direction vector, y image axis in RAS+ coordinates system.

PySisyphe uses RAS+ world coordinates system convention (as MNI, Nibabel, Dipy…) with origin to corner of the voxel:

x, direction [1.0, 0.0, 0.0]: left(-) to right(+)

y, direction [0.0, 1.0, 0.0]: posterior(-) to anterior(+)

z, direction [0.0, 0.0, 1.0]: inferior(-) to superior(+)

Directions is a list of 9 float, 3 vectors of 3 floats:

First vector, x-axis image direction, [1.0, 0.0, 0.0] (RAS+ default)

Second vector, y-axis image direction, [0.0, 1.0, 0.0] (RAS+ default)

Third vector, z-axis image direction, [0.0, 0.0, 1.0] (RAS+ default)

Returns

tuple[float, float, float]: vector of the y-axis image

getSize() → list[int] | tuple[int, int, int]

Get image size, i.e. voxel count in each dimension.

Returns

tuple[int, int, int]: image size in x, y, z dimensions

Get skewness of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int, index of the component to process
None, all components processed. In this case, the method returns a list of skewness values of each

component of the multi-component image

Returns

float | list[float]

skewness value for single-component image
skewness value for multi-component image, if component index c is int (default 0)
list of kurtosis values of each component of a multi-component image, if component index c is None

getSpacing() → list[float] | tuple[float, float, float]

Get voxel size (mm) in each dimension.

Returns

tuple[float, float, float]: voxel spacing in x, y, z

getStandardizeIntensity(method: str = 'norm') → SisypheImage

Get an intensity normalized copy of the current SisypheImage instance.

Parameters

methodstr

‘norm’, standardize the intensity as zscore (i.e. zero mean, [-std, +std] mapped to [-1, 1]
‘rescale’, standardize the intensity values to the range [0, 1]

Returns

SisypheImage
standardized image

getStatistics(mask: str | SisypheImage | SisypheBinaryImage | None = None, c: int = 0) → dict[str, float]

Get descriptive statistics of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint

component index

Returns

dict[str, float]

key/value:

‘mean’: mean
‘std’: standard deviation
‘min’: minimum
‘p25’: first quartile
‘median’: median
‘p75’: third quartile
‘max’: maximum
‘kurtosis’: kurtosis
‘skewness’: skewness

Get standard deviation of image. Calculation is performed on the entire image (mask parameter = None) or in a mask.

Parameters

maskstr | SisypheImage | SisypheBinaryImage | None

str, automatic thresholding algorithm used for background segmentation: ‘mean’, ‘otsu’, ‘huang’, ‘renyi’, ‘yen’, ‘li’, ‘shanbhag’, ‘triangle’, ‘intermodes’, ‘maximumentropy’, ‘kittler’, ‘isodata’ or ‘moments’
SisypheBinaryImage, binary mask image
None, no mask, calculation is performed on the entire image

cint | None

parameter only used for multi-component image
int index of the component to process
None, all components processed. In this case, the method returns a list of standard deviation values of

each component of the multi-component image

Returns

float | list[float]

standard deviation value for single-component image
standard deviation value for multi-component image, if component index c is int (default 0)
list of standard deviation values of each component of a multi-component image, if component index c is None

getThirdVectorDirection() → list[float] | tuple[float, float, float]

Get third direction vector, z image axis in RAS+ coordinates system.

PySisyphe uses RAS+ world coordinates system convention (as MNI, Nibabel, Dipy…) with origin to corner of the voxel:

x, direction [1.0, 0.0, 0.0]: left(-) to right(+)

y, direction [0.0, 1.0, 0.0]: posterior(-) to anterior(+)

z, direction [0.0, 0.0, 1.0]: inferior(-) to superior(+)

Directions is a list of 9 float, 3 vectors of 3 floats:

First vector, x-axis image direction, [1.0, 0.0, 0.0] (RAS+ default)

Second vector, y-axis image direction, [0.0, 1.0, 0.0] (RAS+ default)

Third vector, z-axis image direction, [0.0, 0.0, 1.0] (RAS+ default)

Returns

tuple[float, float, float]: vector of the z-axis image

getTruncateIntensity(centile: int = 1, outputrange: tuple[float, float] | None = None) → SisypheImage

Get intensity truncated copy of the current SisypheImage instance. Truncate threshold is expressed in percentile (min threshold = centile, max threshold = 100 - centile). The max and min values of the output image are given in the output range parameter. If output range is None, max and min values of the output image are the max and min truncate thresholds.

Parameters

centileint: truncate threshold expressed in percentile
outputrangetuple[float, float] | None: max and min values of the output image

Returns

SisypheImage
truncated intensity image

getVTKImage() → vtkImageData

VTKImageData view (pointer) of the current SisypheImage. Image buffer is shared between VTKImageData and current SisypheImage instances.

Returns

vtk.vtkImageData: shallow copy of image

getVoxelCoordinatesFromWorldCoordinates(p: list[float] | tuple[float, float, float]) → list[float] | tuple[float, float, float]

Convert world coordinates to voxel grid coordinates.

Parameters

plist[float, float, float] | tuple[float, float, float]: world coordinates

Returns

tuple[float, float, float]: voxel coordinates (image reference)

getVoxelVolume() → float

Get voxel volume in mm3.

Returns

float: voxel volume in mm3

getWidth() → int

Get image size in x.

Returns

int: x-axis image size

getWorldCoordinatesFromVoxelCoordinates(p: list[float] | tuple[float, float, float]) → list[float] | tuple[float, float, float]

Convert voxel grid coordinates to world coordinates.

Parameters

plist[float, float, float] | tuple[float, float, float]: voxel coordinates (image reference)

Returns

tuple[float, float, float]: world coordinates

hasITKImage() → bool

Checks whether itk.Image attribute is available.

Returns

bool: True if itk.Image attribute is available (not None)

Compare field of view between current SisypheImage instance and other image.

Parameters

imgSisypheImage | SimpleITK.Image | itk.Image | vtk.vtkImageData | ants.core.ANTsImage | tuple[float, float, float] | list[float, float, float]: get FOV value from image attributes (SisypheImage | SimpleITK.Image | itk.Image | vtk.vtkImageData | ants.core.ANTsImage) or directly from a tuple or a list of three values
decimalsint: Number of decimal places to round to (default 2)

Returns

bool: True if same field of view

Compare origin between the current SisypheImage instance and other image.

Parameters

get origin from image attribute (SisypheImage, SimpleITK.Image, itk.Image, vtk.vtkImageData, ants.core.ANTsImage),
or from a tuple, a list, a numpy.ndarray of three values,
or from template tags ‘ICBM152’, ‘ICBM452’, ‘ATROPOS’, ‘SRI24’

Returns

bool: True if same origin

Compare size between the current SisypheImage instance and other image.

Parameters

imgSisypheImage | SimpleITK.Image | itk.Image | vtk.vtkImageData | ants.core.ANTsImage | numpy.ndarray | tuple[int, int, int] | list[int, int, int]: image to compare

Returns

bool: True if same size

hasSameSpacing(img: SisypheImage | Image | ndarray | list[float] | tuple[float, float, float], decimals: int = 2)

Compare spacing between the current SisypheImage instance and other image.

Parameters

imgSisypheImage | SimpleITK.Image | itk.Image | vtk.vtkImageData | ants.core.ANTsImage | tuple[float, float, float] | list[float, float, float]: get spacing from image attribute (SisypheImage | SimpleITK.Image | itk.Image | vtk.vtkImageData | ants.core.ANTsImage) or directly from a tuple or a list of three values
decimalsint: Number of decimal to round to (default 2)

Returns

bool: True if same spacing

isAnisotropic(tol: float = 2.0) → bool

Check whether voxel is anisotropic, i.e. not same spacing in each dimension A tolerance is applied: max(spacing) / min(spacing) > tol

Parameters

tolfloat: tolerance (default 2.0)

Returns

bool: True if max(spacing) / min(spacing) > tol

isDefaultOrigin() → bool

Check whether geometrical reference origin is (0.0, 0.0, 0.0).

Returns

bool: True if world coordinates origin is (0.0, 0.0, 0.0)

isEmpty() → bool

Check whether image buffer is allocated.

Returns

bool: True if image buffer is allocated

isEmptyArray() → bool

Check whether image is empty i.e. all scalar values in the image array are 0.0.

Returns

bool: True if image is empty

isFloat32Datatype() → bool

Check whether datatype is float32.

Returns

bool: True if float32 datatype

isFloat64Datatype() → bool

Check whether datatype is float64.

Returns

bool: True if float64 datatype

isFloatDatatype() → bool

Check whether datatype is float (‘float32’ or ‘float64’).

Returns

bool: True if float datatype

isInt16Datatype() → bool

Check whether datatype is int16.

Returns

bool: True if int16 datatype

isInt32Datatype() → bool

Check whether datatype is int32.

Returns

bool: True if int32 datatype

isInt64Datatype() → bool

Check whether datatype is int64.

Returns

bool: True if int64 datatype

isInt8Datatype() → bool

Check whether datatype is int8.

Returns

bool: True if int8 datatype

isIntegerDatatype() → bool

Check whether datatype is integer (‘uint8’, ‘int8’, ‘uint16’, ‘int16’, ‘uint32’, ‘int32’, ‘uint64’ or ‘int64’).

Returns

bool: True if integer datatype

isIsotropic(tol: float = 2.0) → bool

Check whether voxel is isotropic, i.e. same spacing in each dimension. A tolerance is applied: max(spacing) / min(spacing) <= tol

Parameters

tolfloat: tolerance (default 2.0)

Returns

bool: True if max(spacing) / min(spacing) < tol

isMulticomponent() → bool

Check whether image is multicomponent (array element is a vector and not a scalar value).

Returns

bool: True if multicomponent image

isThickAnisotropic(tol: float = 2.0, thickness: float = 3.0) → bool

Check if the voxel is anisotropic and thicker than a given slice thickness in mm. A tolerance is applied for anisotropy: max(spacing) / min(spacing) > tol.

Parameters

tolfloat: tolerance (default 2.0)
thicknessfloat: slice thickness threshold (default 3.0 mm)

Returns

bool: True if max(spacing) / min(spacing) > tol and max(spacing) >= thickness

isUInt16Datatype() → bool

Check whether datatype is uint16.

Returns

bool: True if uint16 datatype

isUInt32Datatype() → bool

Check whether datatype is uint32.

Returns

bool: True if uint32 datatype

isUInt64Datatype() → bool

Check whether datatype is uint64.

Returns

bool: True if uint64 datatype

isUInt8Datatype() → bool

Check whether datatype is uint8.

Returns

bool: True if uint8 datatype

loadFromBrainVoyagerVMR(filename: str) → dict

Load the current SisypheImage instance from a BrainVoyager (.vmr) file.

Parameters

filenamestr: BrainVoyager file name

loadFromFreeSurferMGH(filename: str) → None

Load the current SisypheImage instance from a FreeSurfer MGH (.mgh, .mgz) file.

Parameters

filenamestr: FreeSurfer file name

loadFromMINC(filename: str) → None

Load the current SisypheImage instance from a Minc (.minc) file.

Parameters

filenamestr: Minc file name

loadFromNIFTI(filename: str, reorient: bool = True) → None

Load the current SisypheImage instance from a Nifti (.nii) file.

Parameters

filenamestr: Nifti file name
reorientbool: conversion to RAS+ world coordinates convention if True (default True)

loadFromNRRD(filename: str) → None

Load the current SisypheImage instance from a Nrrd (.nrrd) file.

Parameters

filenamestr: Nrrd file name

loadFromNumpy(filename: str) → None

Load the current SisypheImage instance from a numpy (.npy) file.

Parameters

filenamestr: numpy file name

loadFromSisyphe(filename: str) → dict

Load the current SisypheImage instance from an old Sisyphe (.vol) file.

Parameters

filenamestr: old Sisyphe binary file name

Returns

dict: Old Sisyphe header (see Sisyphe.sisypheImageIO.loadFromSisyphe docstring for keys and values documentation)

loadFromVTK(filename: str) → None

Load the current SisypheImage instance from a VTK (.vtk) file.

Parameters

filenamestr: VTK file name

noiseEstimate() → float

Estimator of the (Gaussian) noise standard deviation.

Returns

float: noise standard deviation

removeNeckSlices(f: float = 1.8) → SisypheImage

Get a new SisypheImage instance, cropped in z. Most sagittal MRI scans have extensive and useless inferior coverage. This function crop MR volume in z direction, removing empty slices and lower slices of neck below foramen magnum.

Parameters

ffloat: multiplicative factor to adjust the neck slice. Lower values remove more slices, higher values keep more slices (between 1.5 and 2.0, default 1.8 close to the foramen magnum for most MR images)

Returns

SisypheImage

replaceNanInfValues(nan: float = 0.0, posinf: float = 0.0, neginf: float = 0.0) → None

Replace NaN and Inf values with a given values.

Parameters

nanfloat, optional: value to replace NaN (default 0.0)
posinffloat, optional: value to replace positive infinity (default 0.0)
neginffloat, optional: value to replace negative infinity (default 0.0)

saveToJSON(filename: str) → None

Save the current SisypheImage instance to a Json (.json) file.

Parameters

filenamestr: Json file name

saveToMINC(filename: str) → None

Save the current SisypheImage instance to a Minc (.minc) file.

Parameters

filenamestr: Minc file name

saveToNIFTI(filename: str, compress: bool = False) → None

Save the current SisypheImage instance to a Nifti (.nii) file.

Parameters

filenamestr: Nifti file name
compressbool: gzip compression if True (default False)

saveToNRRD(filename: str) → None

Save the current SisypheImage instance to a Nrrd (.nrrd) file.

Parameters

filenamestr: Nrrd file name

saveToNumpy(filename: str) → None

Save the current SisypheImage instance to a numpy (.npy) file.

Parameters

filenamestr: numpy file name

saveToVTK(filename: str) → None

Save the current SisypheImage instance to a vtk (.vtk) file.

Parameters

filenamestr: VTK file name

setDefaultOrigin() → None: Set geometrical reference origin coordinates to (0.0, 0.0, 0.0).

setDirections(direction: list | tuple = (1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0)) → None

Set vectors of image axes in RAS+ coordinates system. PySisyphe uses RAS+ world coordinates system convention (as MNI, Nibabel, Dipy…) with origin to corner of the voxel.

x, direction [1.0, 0.0, 0.0]: left(-) to right(+)

y, direction [0.0, 1.0, 0.0]: posterior(-) to anterior(+)

z: direction [0.0, 0.0, 1.0]: inferior(-) to superior(+)

Directions is a list of 9 float, 3 vectors of 3 floats

First vector, x image axis direction, [1.0, 0.0, 0.0] (RAS+ default)

Second vector, y image axis direction, [0.0, 1.0, 0.0] (RAS+ default)

Third vector, y image axis direction, [0.0, 0.0, 1.0] (RAS+ default)

Parameters

direction tuple[float, …]: 9 elements, vectors of image axes

setOrigin(origin: list[float] | tuple[float, float, float] = (0.0, 0.0, 0.0)) → None

Set geometrical reference origin coordinates.

Parameters

originlist[float, float, float] | tuple[float, float, float]: image origin in world coordinates

setOriginToCenter() → None: Set geometrical reference origin coordinates to image center.

setSITKImage(img: Image) → None

Shallow copy of a SimpleITK Image to the current SisypheImage instance. Image buffer is shared between SimpleITK image and current SisypheImage instances.

Parameters

imgSimpleITK.Image: image to copy

setSpacing(sx: float, sy: float, sz: float) → None

Set voxel size (mm) in each dimension.

Parameters

sxfloat: voxel spacing in x
syfloat: voxel spacing in y
szfloat: voxel spacing in x

sliceHeadSurface(orient: str = 'e') → dict[str, list]

Get head surface (number of pixels) in each slice of the current SisypheImage instance.

Parameters

orientstr: slice orientation (‘a’ axial, ‘c’ coronal, ‘s’ sagittal, ‘e’ every orientation default)

Returns

dict[str, list]

keys: ‘a’ axial results, ‘c’ coronal results, ‘s’ sagittal results
dict value is list with four elements
first element = int, index of the slice with the maximum head surface
second element = int, index of the first slice with non-zero head surface
third element = int, index of the last slice with non-zero head surface
fourth element = numpy.ndarray, head surface (number of pixels) of each slice

sliceIsEmpty(index: int | list[int] | tuple[int, ...], orient: int) → bool

Check whether slices are empty i.e. all scalar values in slices are 0.0.

Parameters

indexint | list[int] | tuple[int, …]: slice indices
orientint: slice orientation code (0 axial, 1 coronal, 2 sagittal)

Returns

bool: True if slices are empty

standardizeIntensity(method: str = 'norm') → None

Intensity normalization of the current SisypheImage instance.

Parameters

methodstr

‘norm’, standardize the intensity as zscore (i.e. zero mean, [-std, +std] mapped to [0, 1]
‘rescale’, standardize the intensity values to the range [0, 1]

truncateIntensity(centile: int = 1, outputrange: tuple[float, float] | None = None)

Truncate intensity of the current SisypheImage instance. Truncate threshold is expressed in percentile (min threshold = centile, max threshold = 100 - centile). The max and min values of the output image are given in the output range parameter. If output range is None, max and min values of the output image are the max and min truncate thresholds.

Parameters

centileint: truncate threshold expressed in percentile
outputrangetuple[float, float] | None: max and min values of the output image

Sisyphe.core.sisypheImage.numpyToVtkImageData(np: ndarray) → vtkImageData

numpy ndarray to vtk.vtkImageData conversion.

Parameters

npnumpy.ndarray: numpy array to convert

Returns

vtk.vtkImageData: converted vtkImageData

Sisyphe.core.sisypheImage.simpleITKToVTK(img: Image) → vtkImageData

SimpleITK.Image to vtk.vtkImageData conversion.

Parameters

imgSimpleITK.Image: SimpleITK.Image to convert

Returns

vtk.vtkImageData: converted vtkImageData