intensity_normalization.normalize package

Submodules

intensity_normalization.normalize.base module

Base class for normalization methods.

class intensity_normalization.normalize.base.DirectoryNormalizeCLI

Bases: SampleNormalizeCLIMixin, DirectoryCLI

before_fit(images: collections.abc.Sequence[intnormt.ImageLike], /, masks: collections.abc.Sequence[intnormt.ImageLike] | None = None, *, modality: intnormt.Modality = Modality.T1, **kwargs: Any) → tuple[ImageSeq, MaskSeqOrNone]

fit(images: Sequence[ImageLike], /, masks: Sequence[ImageLike] | None = None, *, modality: Modality = Modality.T1, **kwargs: Any) → None

fit_from_directories(image_dir: intnormt.PathLike, /, mask_dir: intnormt.PathLike | None = None, *, modality: intnormt.Modality = Modality.T1, ext: str = 'nii*', return_normalized_and_masks: bool = False, **kwargs: Any) → tuple[ImageSeq, MaskSeqOrNone] | None

class intensity_normalization.normalize.base.LocationScaleCLIMixin(*, norm_value: float = 1.0, **kwargs: Any)

Bases: LocationScaleMixin, NormalizeCLIMixin

classmethod from_argparse_args(args: Namespace, /) → T

classmethod get_parent_parser(desc: str, valid_modalities: frozenset[str] = frozenset({'flair', 'md', 'other', 'pd', 't1', 't2'}), **kwargs: Any) → ArgumentParser

class intensity_normalization.normalize.base.SingleImageNormalizeCLI

Bases: NormalizeCLIMixin, SingleImageCLI

call_from_argparse_args(args: Namespace, /, **kwargs: Any) → None

plot_histogram_from_args(args: argparse.Namespace, /, normalized: intnormt.ImageLike, mask: intnormt.ImageLike | None = None) → None

intensity_normalization.normalize.fcm module

Fuzzy C-Means-based tissue mean normalization.

class intensity_normalization.normalize.fcm.FCMNormalize(*, norm_value: float = 1.0, tissue_type: TissueType = TissueType.WM, **kwargs: Any)

Bases: LocationScaleCLIMixin, SingleImageNormalizeCLI

static add_method_specific_arguments(parent_parser: ArgumentParser) → ArgumentParser

calculate_location(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

calculate_scale(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

call_from_argparse_args(args: Namespace, /, **kwargs: Any) → None

static description() → str

classmethod from_argparse_args(args: Namespace, /) → FCMNormalize

static fullname() → str

classmethod get_parent_parser(desc: str, valid_modalities: frozenset[str] = frozenset({'flair', 'md', 'other', 'pd', 't1', 't2'}), **kwargs: Any) → ArgumentParser

property is_fit: bool

static name() → str

save_additional_info(args: Namespace, **kwargs: Any) → None

intensity_normalization.normalize.kde module

Kernel density estimation-based tissue mode normalization.

class intensity_normalization.normalize.kde.KDENormalize(norm_value: float = 1.0, **kwargs: Any)

Bases: LocationScaleCLIMixin, SingleImageNormalizeCLI

calculate_location(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

calculate_scale(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

static description() → str

static fullname() → str

static name() → str

intensity_normalization.normalize.lsq module

Least-squares fit tissue means of a set of images.

class intensity_normalization.normalize.lsq.LeastSquaresNormalize(*, norm_value: float = 1.0, **kwargs: Any)

Bases: LocationScaleCLIMixin, DirectoryNormalizeCLI

static add_method_specific_arguments(parent_parser: ArgumentParser) → ArgumentParser

calculate_location(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

calculate_scale(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

call_from_argparse_args(args: Namespace, /, **kwargs: Any) → None

static description() → str

classmethod from_argparse_args(args: Namespace, /) → LeastSquaresNormalize

static fullname() → str

classmethod get_parent_parser(desc: str, valid_modalities: frozenset[str] = frozenset({'flair', 'md', 'other', 'pd', 't1', 't2'}), **kwargs: Any) → ArgumentParser

load_standard_tissue_means(filename: str | PathLike, /) → None

static name() → str

save_additional_info(args: Namespace, **kwargs: Any) → None

save_standard_tissue_means(filename: str | PathLike, /) → None

scaling_factor(tissue_means: ndarray[Any, dtype[ScalarType]]) → float

static tissue_means(image: ImageLike, /, tissue_membership: ImageLike) → ndarray[Any, dtype[ScalarType]]

intensity_normalization.normalize.nyul module

Nyul & Udupa piecewise linear histogram matching normalization.

class intensity_normalization.normalize.nyul.NyulNormalize(*, output_min_value: float = 1.0, output_max_value: float = 100.0, min_percentile: float = 1.0, max_percentile: float = 99.0, percentile_after_min: float = 10.0, percentile_before_max: float = 90.0, percentile_step: float = 10.0)

Bases: DirectoryNormalizeCLI

static add_method_specific_arguments(parent_parser: ArgumentParser) → ArgumentParser

call_from_argparse_args(args: Namespace, /, **kwargs: Any) → None

static description() → str

classmethod from_argparse_args(args: Namespace, /) → NyulNormalize

static fullname() → str

get_landmarks(image: ImageLike, /) → ndarray[Any, dtype[ScalarType]]

load_standard_histogram(filename: str | PathLike) → None

static name() → str

normalize_image(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → intnormt.ImageLike

property percentiles: ndarray[Any, dtype[ScalarType]]

save_additional_info(args: Namespace, **kwargs: Any) → None

save_standard_histogram(filename: str | PathLike) → None

intensity_normalization.normalize.ravel module

RAVEL normalization (WhiteStripe then CSF correction).

class intensity_normalization.normalize.ravel.RavelNormalize(*, membership_threshold: float = 0.99, register: bool = True, num_unwanted_factors: int = 1, sparse_svd: bool = False, whitestripe_kwargs: dict[str, Any] | None = None, quantile_to_label_csf: float = 1.0, masks_are_csf: bool = False)

Bases: DirectoryNormalizeCLI

static add_method_specific_arguments(parent_parser: ArgumentParser) → ArgumentParser

create_image_matrix_and_control_voxels(images: collections.abc.Sequence[intnormt.ImageLike], /, masks: collections.abc.Sequence[intnormt.ImageLike] | None = None, *, modality: intnormt.Modality = Modality.T1) → tuple[npt.NDArray, npt.NDArray]

creates a matrix of images; rows correspond to voxels, columns are images

Parameters:

• images – list of MR images of interest

• masks – list of corresponding brain masks

• modality – modality of the set of images (e.g., t1)

Returns:

rows are voxels, columns are images control_voxels: rows are csf intersection voxels, columns are images

Return type:

image_matrix

static description() → str

estimate_unwanted_factors(control_voxels: ndarray[Any, dtype[ScalarType]]) → ndarray[Any, dtype[ScalarType]]

classmethod from_argparse_args(args: Namespace, /) → RavelNormalize

static fullname() → str

static name() → str

normalize_image(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → intnormt.ImageLike

process_directories(image_dir: intnormt.PathLike, /, mask_dir: intnormt.PathLike | None = None, *, modality: intnormt.Modality = Modality.T1, ext: str = 'nii*', return_normalized_and_masks: bool = False, **kwargs: Any) → tuple[list[mioi.Image], list[mioi.Image] | None] | None

save_additional_info(args: Namespace, **kwargs: Any) → None

set_template(template: intnormt.ImageLike | ants.ANTsImage) → None

set_template_mask(template_mask: intnormt.ImageLike | ants.ANTsImage | None) → None

teardown() → None

property template: ants.ANTsImage | None

property template_mask: ants.ANTsImage | None

use_mni_as_template() → None

intensity_normalization.normalize.whitestripe module

WhiteStripe (normal-appearing white matter mean & standard deviation) normalization.

class intensity_normalization.normalize.whitestripe.WhiteStripeNormalize(*, norm_value: float = 1.0, width: float = 0.05, width_l: float | None = None, width_u: float | None = None, **kwargs: Any)

Bases: LocationScaleCLIMixin, SingleImageNormalizeCLI

static add_method_specific_arguments(parent_parser: ArgumentParser) → ArgumentParser

calculate_location(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

calculate_scale(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

static description() → str

classmethod from_argparse_args(args: Namespace, /) → WhiteStripeNormalize

static fullname() → str

static name() → str

plot_histogram_from_args(args: argparse.Namespace, /, normalized: intnormt.ImageLike, mask: intnormt.ImageLike | None = None) → None

setup(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → None

teardown() → None

intensity_normalization.normalize.zscore module

Z-score normalize image (voxel-wise subtract mean, divide by standard deviation).

class intensity_normalization.normalize.zscore.ZScoreNormalize(*, norm_value: float = 1.0, **kwargs: Any)

Bases: LocationScaleCLIMixin, SingleImageNormalizeCLI

calculate_location(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

calculate_scale(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → float

static description() → str

static fullname() → str

static name() → str

plot_histogram_from_args(args: argparse.Namespace, /, normalized: intnormt.ImageLike, mask: intnormt.ImageLike | None = None) → None

setup(image: intnormt.ImageLike, /, mask: intnormt.ImageLike | None = None, *, modality: intnormt.Modality = Modality.T1) → None

teardown() → None

Module contents