matchpoint.core.MatchPoint#

class matchpoint.core.MatchPoint(source=None, destination=None, method=None, transformation_type='affine', initial_transformation=None, transformation=None, transformation_inverse=None, source_name='source', destination_name='destination', source_unit=None, destination_unit=None, destination_distance_threshold=0, name=None)#

Mapping class to find, improve, store and use a mapping between a source point set and a destination point set.

source#

Source coordinates.

Type:: Nx2 numpy.ndarray

destination#

Destination coordinates.

Type:: Nx2 numpy.ndarray

method#

Method for finding the transformation between source and destination.

Type:: str

transformation_type#

Type of transformation used. Choose from: affine (default), similarity or polynomial.

Type:: str

initial_transformation#

Initial transformation used as starting point by the perform_mapping method. Can be given as dictionary specifying translation, scale, rotation and shear.

Type:: skimage.transform._geometric.GeometricTransform or dict

transformation#

Transformation relating the source and destionaton point sets. Can be given as dictionary specifying translation, scale, rotation and shear.

Type:: skimage.transform._geometric.GeometricTransform or dict

transformation_inverse#

Inverse transformation. Can be given as dictionary specifying translation, scale, rotation and shear.

Type:: skimage.transform._geometric.GeometricTransform or dict

source_name#

Name of the source. Default is ‘source’.

Type:: str

destination_name#

Name of the source. Default is ‘destination_name’.

Type:: str

source_unit#

Unit of the source coordinates.

Type:: str

destination_unit#

Unit of the destination coordinates.

Type:: str

destination_distance_threshold#

Distance threshold in destination units for determining matched pairs between source and destination.

Type:: float

name#

Name of the mapping, used as the default filename when saving the mapping object.

Type:: str

label#

Additional attribute to store information.

Type:: str

matched_pairs#

Array with the indices of source points and destination points that are matched.

Type:: Nx2 numpy.ndarray

save_path#

Path to the folder where the MatchPoint objects is or should be saved.

Type:: pathlib2.Path

__init__(source=None, destination=None, method=None, transformation_type='affine', initial_transformation=None, transformation=None, transformation_inverse=None, source_name='source', destination_name='destination', source_unit=None, destination_unit=None, destination_distance_threshold=0, name=None)#

Initialize a MatchPoint object

Parameters:

source (Nx2 numpy.ndarray) – Source coordinates.
destination (Nx2 numpy.ndarray) – Destination coordinates.
method (str, optional) – Method for finding the transformation between source and destination.
transformation_type (str) – Type of transformation used. Choose from: affine (default), similarity or polynomial.
initial_transformation (skimage.transform._geometric.GeometricTransform or dict, optional) – Initial transformation used as starting point by the perform_mapping method. Can be given as dictionary specifying translation, scale, rotation and shear.
transformation (skimage.transform._geometric.GeometricTransform or dict, optional) – Transformation relating the source and destionaton point sets. Can be given as dictionary specifying translation, scale, rotation and shear.
transformation_inverse (skimage.transform._geometric.GeometricTransform or dict, optional) – Inverse transformation. Can be given as dictionary specifying translation, scale, rotation and shear.
source_name (str, optional) – Name of the source. Default is ‘source’.
destination_name (str, optional) – Name of the source. Default is ‘destination_name’.
source_unit (str, optional) – Unit of the source coordinates.
destination_unit (str, optional) – Unit of the destination coordinates.
destination_distance_threshold (float) – Distance threshold in destination units for determining matched pairs between source and destination.
name (str) – Name of the mapping, used as the default filename when saving the mapping object.

Methods

`Ripleys_K`([crop, space])
`Ripleys_L_minus_d`([crop, space, plot])
`Ripleys_L_minus_d_max`([crop, space])
`__init__`([source, destination, method, ...])	Initialize a MatchPoint object
`calculate_inverse_transformation`()	Calculates the reverse transformation from the forward transformation.
`cross_correlation`([peak_detection, ...])	Perform cross correlation on synthesized images of the two datasets.
`density_destination`([crop, space])	Destination density estimation from the number of points and the area contained within the convex hull.
`density_source`([crop, space])	Source density estimation from the number of points and the area contained within the convex hull.
`determine_matched_pairs`([...])	Find pairs of source and destination points that closer than the distance threshold and are singly matched, i.e. each source point only has one destination point within the threshold and vice versa.
`direct_match`([transformation_type])	Find transformation from source to destination points by matching based on the point order
`distance_matrix`([crop, space, margin, ...])	Distance matrix for all point combinations of the source and destination.
`find_distance_threshold`([method])	Find distance optimal distance threshold and automatically sets the destination distance threshold.
`fraction_of_destination_matched`([crop])	Fraction of the destination points that is matched to a source point
`fraction_of_source_matched`([crop])	Fraction of the source points that is matched to a destination point
`geometric_hashing`([method, tuple_size, ...])	Perform mapping by geometric hashing.
`get_destination`([crop, space, margin])	Getter for the destination point set.
`get_destination_area`([crop, space])	Get the area of the destination point set, i.e. within the destination point vertices.
`get_destination_vertices`([crop, space, margin])	Getter for vertices of the destination point set.
`get_overlap_vertices`([space])	Getter for vertices of the overlap between the source and destination point sets.
`get_source`([crop, space, margin])	Getter for the source point set.
`get_source_area`([crop, space])	Get the area of the source point set, i.e. within the source point vertices.
`get_source_vertices`([crop, space, margin])	Getter for vertices of the source point set.
`get_transformation_direction`(direction)	Get inverse parameter based on direction
`get_unit`(space)	Get the unit of the source or destination point set.
`get_unit_label`(space)	Get the string to add the axis label during plotting.
`iterative_closest_point`([...])	Find transformation from source to destination points using an iterative closest point algorithm
`kernel_correlation`([bounds, sigma, crop, plot])	Find transformation from source to destination points using an kernel correlation algorithm.
`kernel_correlation_score`([sigma, crop, ...])	Determine the kernel correlation score for the current transformation.
`load`(filepath)	Load a saved MatchPoint object
`nearest_neighbour_match`([...])	Find transformation from source to destination points by matching nearest neighbours.
`number_of_matches_for_source_and_destination`([...])	Gives the number of source and destination points that respectively have a specified number of matches with the destination and source point sets.
`number_of_single_matches_for_radii`(...)	The number of singly-matched pairs for the provided distance thresholds.
`pair_coordinates`([point_set_name, space])	Get the coordinates of the paired source or destination points.
`pair_distances`([space, show])	Distances between the paired source and destination points.
`perform_mapping`([method])	Find transformation from source to destination points using one of the mapping methods
`save`([save_path, filetype])	Save the current mapping in a file, so that it can be opened later.
`set_correlation_peak_coordinates`(...)	Manually set the peak coordinates after cross-cross correlation.
`show`([axis, show_source, show_destination, ...])	Show the mapping.
`show_destination`(**kwargs)	Plot the destination point set.
`show_mapping_transformation`(args, *kwargs)	Show the mapping.
`show_outline`([inverse, source_colour, ...])	Show the outline or border of the point sets.
`show_source`(**kwargs)	Plot the source point set.
`simulate`([number_of_points, transformation, ...])	Simulate a point set with a specific transformation and return as a MatchPoint object.
`single_match_optimization`([maximum_radius, ...])	Find the distance threshold with the highest number of singly matched pairs.
`transform_coordinates`(coordinates[, ...])	Transform coordinates using the object's transformation.
`transform_image`(image[, inverse, direction])	Transform an image using the objects transformation.
`transformation_is_similar_to_correct_transformation`(...)

Attributes

Ripleys_K(crop=True, space='destination')#

Ripleys_L_minus_d(crop=True, space='destination', plot=False)#

Ripleys_L_minus_d_max(crop=True, space='destination')#

calculate_inverse_transformation()#

Calculates the reverse transformation from the forward transformation.

Sets the object’s transformation_inverse attribute.

Note

This is only possible for linear transformations.

cross_correlation(peak_detection='auto', kernel_size=7, gaussian_sigma=1, divider=5, crop=False, space='destination', normalize=False, subtract_background=True, plot=False, axes=None)#

Perform cross correlation on synthesized images of the two datasets.

The current transformation is used as a starting point. Because cross-correlation only varies translation, other transformation parameters such rotation and scale should already be correct. The found transformation is stored in the object’s transformation attribute.

Parameters:

peak_detection (str) – If ‘auto’: the cross-correlation peak is automatically determined by finding the most intense pixel in the cross-correlation image. If ‘manual’: cross-correlation image is created, but the peak has to be determined manually. The found pixel coordinates of the peak in the image can be set using the method ‘set_correlation_peak_coordinates’, the correct transformation will then be calculated.
kernel_size (int) – The size in pixels of the gaussian kernel. The kernel will have shape (kernel_size x kernel_size).
gaussian_sigma (float) – The standard deviation of the gaussian kernel. The same standard deviation is used for x and y.
divider (int or float) – The number by which the point set coordinates are divided to obtain pixel indices. (Translating the lowest coordinates to the pixel (0,0) is done automatically.)
crop (int or float, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
space (str, optional) – In which coordinate space to perform the correlation. Either ‘source’ or ‘destination’.
normalize (bool, optional) – If True, then calculate the zero-normalized cross correlation. As it this is not done in Fourier space it is (likely) slower than regular cross correlation. If False: no normalization is performed.
subtract_background (bool or str , optional) – If True or ‘expected_signal’: the expected signal based on the intensity at each location is calculated and subtracted. If ‘minimum_filter’: a minimum filter of two times the kernel size is applied and the resulting image is subtracted. If ‘median_filter’: a median filter of two times the kernel size is applied and the resulting image is subtracted. If ‘expected_signal_rough’: the expected signal is estimated based on the overlap of the images and their average values, the resulting image is subtracted. If False: no background subtraction is performed.
plot (bool) – If True: plots of the synthetic images and the cross-correlation image are shown.
axes (list of matplotlib.axis.Axis) – Three axes to use for plotting.

Note

Automatic background subtraction is performed on the cross-correlation image. This reduces the difference in intensity between the edges and the center that naturally occur when doing cross-correlation, and it thus helps to find the correct peak.

density_destination(crop=False, space='destination')#

Destination density estimation from the number of points and the area contained within the convex hull.

Parameters:: crop (int or float, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.

spacestr, optional: In which coordinate space to obtain the destination density. Either ‘source’ or ‘destination’.

Returns:: density_source (float) – Density of the destination dataset

density_source(crop=False, space='source')#

Source density estimation from the number of points and the area contained within the convex hull.

Parameters:: crop (int or float, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.

spacestr, optional: In which coordinate space to obtain the source density. Either ‘source’ or ‘destination’.

Returns:: density_source (float) – Density of the source dataset

property destination_area#

property destination_cropped#

property destination_cropped_area#

property destination_cropped_vertices#

property destination_distance_threshold#

property destination_to_source#

property destination_vertices#

determine_matched_pairs(distance_threshold=None, point_set_name='all')#

Find pairs of source and destination points that closer than the distance threshold and are singly matched, i.e. each source point only has one destination point within the threshold and vice versa.

Sets the matched_pairs attribute.

Parameters:

distance_threshold (int or float) – Distance threshold for match determination.
point_set_name (str) – If ‘source’ then only the source point should have a single match in the destination point set. the destination point can have multiple matches in the source point set. If ‘destination’ then only the destination point should have a single match in the source point set, the source point can have multiple matches in the destination point set. If ‘all’ then both source and destination should have a single match.

direct_match(transformation_type=None, **kwargs)#

Find transformation from source to destination points by matching based on the point order

The found transformation is stored in the object transformation parameter.

Note

The number and the order of source points should be equal to the number and the order of destination points.

Parameters:

transformation_type (str) – Type of transformation used, either linear or polynomial can be chosen. If not specified the object transformation_type is used.
kwargs – Keyword arguments passed to the direct match function.

distance_matrix(crop=True, space='destination', margin=None, max_distance=None, **kwargs)#

Distance matrix for all point combinations of the source and destination.

Parameters:

crop (int or float, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
space (str, optional) – In which coordinate space to obtain the distances. Either ‘source’ or ‘destination’.
margin (float or tuple of floats) – Margins to exclude for the point sets. A tuple should be given in the format (margin_source, margin_destination). If a single margin is passed, then it will be used for both the source and destination.
max_distance (float) – Maximum distance to include in the distance matrix. If given, then a sparse distance matrix will be returned.
kwargs – Keyword arguments to pass to the scipy.spatial.cKDTree.distance_matrix or scipy.spatial.cKDTree.sparse_distance_matrix function.

Returns:

distance_matrix (MxN numpy.ndarray or scipy.sparse.spmatrix) – Distance matrix.

property file_path#

find_distance_threshold(method='single_match_optimization', **kwargs)#

Find distance optimal distance threshold and automatically sets the destination distance threshold.

Parameters:

method (str) –
Method to use for finding the distance threshold. Choose from:
- ‘single_match_optimization’: Optimizes the number of singly-matched pairs, i.e. the source point having only a single destination point within the distance threshold and the destination point having only a single source point within the distance threshold. See the MatchPoint.single_match_optimization method.
kwargs – Keyword arguments to pass to method.

fraction_of_destination_matched(crop=False)#

Fraction of the destination points that is matched to a source point

Parameters:: crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
Returns:: float – Fraction of destination that is matched

fraction_of_source_matched(crop=False)#

Fraction of the source points that is matched to a destination point

Parameters:: crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
Returns:: float – Fraction of source that is matched

geometric_hashing(method='one_by_one', tuple_size=4, maximum_distance_source=None, maximum_distance_destination=None, **kwargs)#

Perform mapping by geometric hashing.

The found transformation is stored in the object’s transformation attribute.

Parameters:

method (str) – The method to use: - ‘one_by_one’: tuples are generated, matched and tested sequentially. - ‘abundant_transformations’: All tuples are generated and matched. Transformation parameters are determined

for matched tuples and the highest density cluster is determined.
tuple_size (int) – Number of points per tuple.
maximum_distance_source (float) – Maximum distance between the two outermost points in a tuple of the source point set.
maximum_distance_destination (float) – Maximum distance between the two outermost points in a tuple of the destination point set.
kwargs – Keyword arguments to pass to the geometric hashing query function.

get_destination(crop=False, space='destination', margin=None)#

Getter for the destination point set.

Parameters:

crop (bool or str, optional) – If True or ‘destination’, the point set is cropped to the area of the source.
space (str, optional) – In which coordinate space to return the point set. Either ‘source’ or ‘destination’.
margin (float or int, optional) – Margin used for cropping.

Returns:

Nx2 numpy.ndaarray – Destination point set

get_destination_area(crop=False, space='destination')#

Get the area of the destination point set, i.e. within the destination point vertices.

Parameters:

crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
space (str, optional) – In which coordinate space to return the point set. Either ‘source’ or ‘destination’.

Returns:

float – Destination area

get_destination_vertices(crop=False, space='destination', margin=None)#

Getter for vertices of the destination point set.

Parameters:

crop (bool or str, optional) – If True or ‘destination’, the vertices of the overlapping area with the source point set are given.
space (str, optional) – In which coordinate space to return the vertices. Either ‘source’ or ‘destination’.
margin (float or int, optional) – Margin used for cropping.

Returns:

Nx2 numpy.ndaarray – Coordinates of the destination vertices

get_overlap_vertices(space='source')#

Getter for vertices of the overlap between the source and destination point sets.

Parameters:: space (str, optional) – In which coordinate space to return the vertices. Either ‘source’ or ‘destination’.
Returns:: Nx2 numpy.ndaarray – Coordinates of the overlap vertices

get_source(crop=False, space='source', margin=None)#

Getter for the source point set.

Parameters:

crop (bool or str, optional) – If True or ‘source’, the point set is cropped to the area of the destination.
space (str, optional) – In which coordinate space to return the point set. Either ‘source’ or ‘destination’.
margin (float or int, optional) – Margin used for cropping.

Returns:

Nx2 numpy.ndaarray – Source point set

get_source_area(crop=False, space='destination')#

Get the area of the source point set, i.e. within the source point vertices.

Parameters:

crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
space (str, optional) – In which coordinate space to return the point set. Either ‘source’ or ‘destination’.

Returns:

float – Source area

get_source_vertices(crop=False, space='source', margin=None)#

Getter for vertices of the source point set.

Parameters:

crop (bool or str, optional) – If True or ‘source’, the vertices of the overlapping area with the destination point set are given.
space (str, optional) – In which coordinate space to return the vertices. Either ‘source’ or ‘destination’.
margin (float or int, optional) – Margin used for cropping.

Returns:

Nx2 numpy.ndaarray – Coordinates of the source vertices

get_transformation_direction(direction)#

Get inverse parameter based on direction

Parameters:

direction (str) –

The direction of the transformation, either ‘<source_name>2<destination_name>’ or: ‘<destination_name>2<source_name>’. ‘source2destination’ and ‘destination2source’ can also be used.

Returns:

inverse (bool) – Specifier for the use of the forward or inverse transformation.

get_unit(space)#

Get the unit of the source or destination point set.

Parameters:: space (str) – For which coordinate space to obtain the unit. Either ‘source’ or ‘destination’.
Returns:: str – The unit of the point set.

get_unit_label(space)#

Get the string to add the axis label during plotting. This is either the unit between parentheses, or if no unit is set it is an empty string.

Parameters:: space (str) – For which coordinate space to obtain the unit. Either ‘source’ or ‘destination’.
Returns:: str – Unit label

iterative_closest_point(distance_threshold=None, verbose=True, **kwargs)#

Find transformation from source to destination points using an iterative closest point algorithm

In the iterative closest point algorithm, the two-way nearest neighbours are found and these are used to find the most optimal transformation. Subsequently, the source is transformed according to this transformation. This process is repeated until the changes detected are below a tolerance level.

The iterative closest point algorithm can be used in situations when deviations between the two point sets are relatively small.

The found transformation is stored in the object’s transformation attribute.

Note

The current transformation is used as starting point for the algorithm.

Note

The printed error is based on the points selected for matching.

Parameters:

distance_threshold (int or float) – Distance threshold applied to select nearest-neighbours in the final round of icp, i.e. nearest-neighbours with distances smaller than distance_threshold.
kwargs – Keyword arguments passed to the algorithm.

kernel_correlation(bounds=((0.97, 1.02), (- 0.05, 0.05), (- 10, 10), (- 10, 10)), sigma=1, crop=False, plot=False, **kwargs)#

Find transformation from source to destination points using an kernel correlation algorithm.

The found transformation is stored in the object’s transformation attribute.

Note

The found kernel correlation value is stored in the objects mapping statistics with key “kernel_correlation_value”.

Parameters:

bounds (tuple of tuples) – The bounds for variation of the transformation parameters. The number of tuples passed determines the number of varied parameters. - Two tuples indicates variation of the x and y translation. - Three tuples indicate rotation, x-translation and y-translation. - Four tuples indicate scale, rotation, x-translation and y-translation. - Five tuples indicate scale, rotation, shear, x-translation and y-translation. - Six tuples indicate scale-x, scale-y, rotation, shear, x-translation and y-translation. The rotation parameter is given in radians.
sigma (float) – The standard deviation of the Gaussian kernel used.
crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
plot (bool) – Whether to plot the mapping process.
kwargs – Keyword arguments passed to the algorithm.

kernel_correlation_score(sigma=1, crop=False, per_point_pair=False)#

Determine the kernel correlation score for the current transformation.

Parameters:

sigma (float) – The standard deviation of the Gaussian kernel used.
crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
per_point_pair (bool) – If True the kernel correlation score is returned for each point pair in a 2D matrix. If False a summed kernel correlation score is returned.

Returns:

float or NxM numpy.ndarray – Kernel correlation score

classmethod load(filepath)#

Load a saved MatchPoint object

Parameters:: filepath (str or pathlib2.Path) – Filepath to the saved object.
Returns:: mapping (MatchPoint) – Saved MatchPoint object.

nearest_neighbour_match(distance_threshold=None, transformation_type=None, **kwargs)#

Find transformation from source to destination points by matching nearest neighbours.

Two-way nearest neighbours are detected, i.e. the source point should be the nearest neighbour of the destination point and vice versa. Only nearest neighbours closer than the distance threshold are used to find the transformation.

The found transformation is stored in the object transformation parameter.

Note

The current transformation is used as starting point for the algorithm.

Note

The printed error is based on the points selected for matching.

Parameters:

distance_threshold (float) – Distance threshold for nearest neighbour match in destination space. Only nearest neighbours with a distance smaller than the distance threshold are used.
transformation_type (str) – Type of transformation used. For options see MatchPoint.transformation_type. If not specified the object transformation_type is used.
kwargs – Keyword arguments passed to the nearest-neighbour match function.

property number_of_destination_points#

property number_of_matched_points#

number_of_matches_for_source_and_destination(distance_threshold=None, matches_per_point=[0, 1], crop=True)#

Gives the number of source and destination points that respectively have a specified number of matches with the destination and source point sets.

Parameters:

distance_threshold (int, float, list or numpy.ndarray) – Distance threshold(s) to determine the number of matches for. If None (default) then the destination_distance_threshold attribute is used.
matches_per_point (int or list of int) – Number of matches to report
crop (bool or str, optional) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.

Returns:

xarray.DataArray – Number of source or destination points with specific distance threshold

number_of_single_matches_for_radii(distance_thresholds)#

The number of singly-matched pairs for the provided distance thresholds.

Here singly-matched pairs are defined as the source point having only a single destination point within the distance threshold and the destination point having only a single source point within the distance threshold.

Parameters:: distance_thresholds (1D numpy.ndaarray) – Distance thresholds to test.
Returns:: 1D numpy.ndaarray – Number of singly-matched points for each distance threshold.

property number_of_source_points#

pair_coordinates(point_set_name='destination', space='destination')#

Get the coordinates of the paired source or destination points.

Parameters:

point_set_name (str) – Name from the point set to use the coordinates from. Either ‘source’ or ‘destination’ (default).
space (str, optional) – In which coordinate space to return the coordinates. Either ‘source’ or ‘destination’ (default).

Returns:

Nx2 numpy.ndaarray – Coordinates of paired points

pair_distances(space='destination', show=False, **kwargs)#

Distances between the paired source and destination points.

Parameters:

space (str, optional) – In which coordinate space to return the coordinates. Either ‘source’ or ‘destination’ (default).
show (bool) – If True, show a histogram of the distances.
kwargs – Keyword arguments passed to matplotlib.pyplot.histogram.

Returns:

1D numpy.ndaarray – Distances between paired points.

perform_mapping(method=None, **kwargs)#

Find transformation from source to destination points using one of the mapping methods

The starting point for the mapping is the initial_transformation attribute.

Parameters:

method (str) – Mapping method, if not specified the object method is used.
kwargs – Keyword arguments passed to mapping methods.

property rotation#

save(save_path=None, filetype='nc')#

Save the current mapping in a file, so that it can be opened later.

Parameters:

filepath (str or pathlib.Path) – Path to file (including filename)
filetype (str) – Choose classic to export a .coeff file for a linear transformation Choose yml to export all object attributes in a yml text file

property scale#

set_correlation_peak_coordinates(correlation_peak_coordinates)#

Manually set the peak coordinates after cross-cross correlation.

Note

This function can only be used after cross-correlation is run.

Parameters:: correlation_peak_coordinates (list, tuple or 1D numpy.ndarray) – Pixel coordinates of the manually found peak in the cross-correlation image.

property shear#

show(axis=None, show_source=False, show_destination=False, show_transformed_coordinates=True, show_pairs=True, crop=False, inverse=False, source_colour='forestgreen', destination_colour='r', pair_colour='b', use_distance_threshold=False, save=False, save_path=None, show_legend=False, return_plot=False)#

Show the mapping.

Show the source and destination points in a scatter plot.

Parameters:

axis (matplotlib.axis.Axis) – The axis to use for plotting. If no axis is given a new figure and axis are created.
show_source (bool) – If True: the original source point set is shown.
show_destination (bool) – If True: the original destination point set is shown.
show_transformed_coordinates (bool) – If True: the transformed point set is shown. If ‘inverse’ is set to False, this is the transformed source point set. If ‘inverse’ is set to True, this is the transformed destination point set.
show_pairs (bool) – If True: the paired points are indicated in a different colour.
crop (bool or str) – If True: the overlap between the source and destination point set is used. If ‘source’: only the source is cropped to the destination point set. If ‘destination’: only the destination is cropped to the source point set. If False: no cropping is applied.
inverse (bool) – Indicates the direction of the transformation to use. If False: the source is transformed to destination space. If True: the destination is transformed to the source space.
source_colour (str) – Colour used for the source points.
destination_colour (str) – Colour used for the destination points.
pair_colour (str) – Colour used to indicate paired points.
use_distance_threshold (bool) – If True: the open circles of the transformed point set will have the size of the source or destination distance threshold. Note: using this option slows down plotting so it is not recommended for large datasets.
save (bool) – If True: the figure is saved in the save path.
save_path (str or pathlib2.Path) – Folderpath to save the figure to. If no path is given, the object’s save_path is used. For the name of the file, the object’s name attribute is used.
show_legend (bool) – If True: the legend is shown.
return_plot (bool) – If True: the figure and axis are returned by the function.

Returns:

figure (matplotlib.figure.Figure) – Figure of the plot
axis (matplotlib.axes.Axes) – Axis of the plot

show_destination(**kwargs)#

Plot the destination point set.

Parameters:

kwargs – Keyword arguments to pass to the show function. Keyword arguments starting with “show” should not be used.

Returns:

figure (matplotlib.figure.Figure) – Figure of the plot
axis (matplotlib.axes.Axes) – Axis of the plot

show_mapping_transformation(*args, **kwargs)#: Show the mapping. Identical to MatchPoint.show().

show_outline(inverse=False, source_colour='forestgreen', destination_colour='r', axis=None)#

Show the outline or border of the point sets.

Parameters:

inverse (bool) – Indicates the direction of the transformation to use. If False: the source is transformed to destination space. If True: the destination is transformed to the source space.
source_colour (str) – Colour used for the source points.
destination_colour (str) – Colour used for the destination points.
axis (matplotlib.axis.Axis) – The axis to use for plotting. If no axis is given a new figure and axis are created.

Returns:

figure (matplotlib.figure.Figure) – Figure of the plot
axis (matplotlib.axes.Axes) – Axis of the plot

show_source(**kwargs)#

Plot the source point set.

Parameters:

kwargs – Keyword arguments to pass to the show function. Keyword arguments starting with “show” should not be used.

Returns:

figure (matplotlib.figure.Figure) – Figure of the plot
axis (matplotlib.axes.Axes) – Axis of the plot

classmethod simulate(number_of_points=200, transformation=None, bounds=[[0, 0], [256, 512]], crop_bounds=(None, None), fraction_missing=(0.1, 0.1), error_sigma=(0.5, 0.5), shuffle=True, seed=10532, show_correct=True)#

Simulate a point set with a specific transformation and return as a MatchPoint object. It generates an original point set of which the source and destination point sets are subsets.

Parameters:

number_of_points (int) – Number of points in the original point set to simulate, the source and destination point sets are drawn from the original point set.
transformation (skimage.transform._geometric.GeometricTransform) – Transformation to be used between the source and destination point sets. If None, a preset transformation will be used: SimilarityTransform(translation=[256, 10], rotation=1/360*2*np.pi, scale=[0.98, 0.98]).
bounds (2x2 numpy.ndarray or list) – Bounds of the original point set. Structured like coordinates, i.e. columns are x and y dimensions, rows are minimum and maximum values. [[x_min, y_min],[x_max, y_max]]
crop_bounds (tuple of 2x2 numpy.ndarray or tuple of 2x2 list, optional) – Bounds to be used for cropping the overall dataset into the source and destination point set (before the transformation is applied).
fraction_missing (tuple(float, float)) – Fraction of points that is removed from the original point set to obtain the source and destination point sets.
error_sigma (tuple(float, float)) – Standard deviation of the Gaussian error applied to the original point set to obtain the source and destination point sets.
shuffle (bool) – If True (default) then the points in the destination point set will be shuffled. If False the order of the destination points in the source and destination will be identical.
seed (int) – Seed used for random number generator.
show_correct (bool) – If True show the generated dataset with the correct transformation.

Returns:

mapping (MatchPoint) – MatchPoint object with the simulated source and destination point sets. While the “transformation” attribute is set with a unit transformation, mapping has an additional attribute “transformation_correct” containing the correct transformation. The returned mapping also has a “show_correct_mapping_transformation” method, to visualize the correct transformation.

single_match_optimization(maximum_radius=20, number_of_steps=100, plot=True)#

Find the distance threshold with the highest number of singly matched pairs.

Test thresholds between 0 and maximum_radius. Destination_distance_threshold is automatically set to the found value.

Parameters:

maximum_radius (float) – Maximum distance threshold to test.
number_of_steps (int) – Number of steps on the interval
plot (bool) – If True, a histogram of the number of pairs per distance threshold will be plotted

property source_area#

property source_cropped#

property source_cropped_area#

property source_cropped_vertices#

property source_distance_threshold#

property source_in_destination_area#

property source_to_destination#

property source_vertices#

transform_coordinates(coordinates, inverse=False, direction=None)#

Transform coordinates using the object’s transformation.

Parameters:

coordinates (Nx2 numpy.ndarray) – Coordinates to be transformed
inverse (bool) – If True: the inverse transformation will be used (i.e. from destination to source)
direction (str) – Another way of specifying the direction of the transformation, choose ‘source2destination’, ‘destination2source’ or use the object’s source and destination names instead of ‘source’ and ‘destination’.

Returns:

Nx2 numpy.ndarray – Transformed coordinates

transform_image(image, inverse=False, direction=None)#

Transform an image using the objects transformation.

Parameters:

image (NxM numpy.ndarray) – Image to be transformed.
inverse (bool) – If True: the inverse transformation will be used (i.e. from destination to source)
direction (str) – Another way of specifying the direction of the transformation, choose ‘source2destination’, ‘destination2source’ or use the object’s source and destination names instead of ‘source’ and ‘destination’.

Returns:

NxM numpy.ndarray – Transformed image

property transformation#

property transformation_inverse#

transformation_is_similar_to_correct_transformation(**kwargs)#

property transformation_type#

transformation_types = {'affine': <class 'skimage.transform._geometric.AffineTransform'>, 'linear': <class 'skimage.transform._geometric.AffineTransform'>, 'nonlinear': <class 'matchpoint.polywarp.PolywarpTransform'>, 'polynomial': <class 'matchpoint.polynomial.PolynomialTransform'>, 'similarity': <class 'skimage.transform._geometric.SimilarityTransform'>}#

property translation#

`destination_area`
`destination_cropped`
`destination_cropped_area`
`destination_cropped_vertices`
`destination_distance_threshold`
`destination_to_source`
`destination_vertices`
`file_path`
`number_of_destination_points`
`number_of_matched_points`
`number_of_source_points`
`rotation`
`scale`
`shear`
`source_area`
`source_cropped`
`source_cropped_area`
`source_cropped_vertices`
`source_distance_threshold`
`source_in_destination_area`
`source_to_destination`
`source_vertices`
`transformation`
`transformation_inverse`
`transformation_type`
`transformation_types`
`translation`