import abc
from collections import OrderedDict, MutableMapping
import numpy as np
from menpo.base import Copyable
from menpo.transform.base import Transformable
from menpo.visualize.base import Viewable
[docs]class Landmarkable(Copyable):
r"""
Abstract interface for object that can have landmarks attached to them.
Landmarkable objects have a public dictionary of landmarks which are
managed by a :map:`LandmarkManager`. This means that
different sets of landmarks can be attached to the same object.
Landmarks can be N-dimensional and are expected to be some
subclass of :map:`PointCloud`. These landmarks
are wrapped inside a :map:`LandmarkGroup` object that performs
useful tasks like label filtering and viewing.
"""
__metaclass__ = abc.ABCMeta
def __init__(self):
self._landmarks = None
@abc.abstractproperty
def n_dims(self):
"""
The total number of dimensions.
:type: `int`
"""
pass
@property
def landmarks(self):
"""
The landmarks object.
:type: :map:`LandmarkManager`
"""
if self._landmarks is None:
self._landmarks = LandmarkManager()
return self._landmarks
@property
def has_landmarks(self):
"""
Whether the object has landmarks.
:type: `bool`
"""
return self._landmarks is not None
@landmarks.setter
def landmarks(self, value):
"""
Landmarks setter.
Parameters
----------
value : :map:`LandmarkManager`
The landmarks to set.
"""
# firstly, make sure the dim is correct. Note that the dim can be None
lm_n_dims = value.n_dims
if lm_n_dims is not None and lm_n_dims != self.n_dims:
raise ValueError(
"Trying to set {}D landmarks on a "
"{}D object".format(value.n_dims, self.n_dims))
self._landmarks = value.copy()
@property
def n_landmark_groups(self):
r"""
The number of landmark groups on this object.
:type: `int`
"""
return self.landmarks.n_groups
[docs]class LandmarkManager(MutableMapping, Transformable):
"""Store for :map:`LandmarkGroup` instances associated with an object
Every :map:`Landmarkable` instance has an instance of this class available
at the ``.landmarks`` property. It is through this class that all access
to landmarks attached to instances is handled. In general the
:map:`LandmarkManager` provides a dictionary-like interface for storing
landmarks. :map:`LandmarkGroup` instances are stored under string keys -
these keys are refereed to as the **group name**. A special case is
where there is a single unambiguous :map:`LandmarkGroup` attached to a
:map:`LandmarkManager` - in this case ``None`` can be used as a key to
access the sole group.
Note that all landmarks stored on a :map:`Landmarkable` in it's attached
:map:`LandmarkManager` are automatically transformed and copied with their
parent object.
"""
def __init__(self):
super(LandmarkManager, self).__init__()
self._landmark_groups = {}
@property
def n_dims(self):
"""
The total number of dimensions.
:type: `int`
"""
if self.n_groups != 0:
# Python version independent way of getting the first value
for v in self._landmark_groups.values():
return v.n_dims
else:
return None
[docs] def copy(self):
r"""
Generate an efficient copy of this :map:`LandmarkManager`.
Returns
-------
``type(self)``
A copy of this object
"""
# do a normal copy. The dict will be shallow copied - rectify that here
new = Copyable.copy(self)
for k, v in new._landmark_groups.items():
new._landmark_groups[k] = v.copy()
return new
def __iter__(self):
"""
Iterate over the internal landmark group dictionary
"""
return iter(self._landmark_groups)
def __setitem__(self, group, value):
"""
Sets a new landmark group for the given label. This can be set using
an existing landmark group, or using a PointCloud. Existing landmark
groups will have their target reset. If a PointCloud is provided then
all landmarks belong to a single label `all`.
Parameters
----------
group : `string`
Label of new group.
value : :map:`LandmarkGroup` or :map:`PointCloud`
The new landmark group to set.
Raises
------
DimensionalityError
If the landmarks and the shape are not of the same dimensionality.
"""
from menpo.shape import PointCloud
# firstly, make sure the dim is correct
n_dims = self.n_dims
if n_dims is not None and value.n_dims != n_dims:
raise ValueError(
"Trying to set {}D landmarks on a "
"{}D LandmarkManager".format(value.n_dims, self.n_dims))
if isinstance(value, PointCloud):
# Copy the PointCloud so that we take ownership of the memory
lmark_group = LandmarkGroup(
value,
OrderedDict([('all', np.ones(value.n_points, dtype=np.bool))]))
elif isinstance(value, LandmarkGroup):
# Copy the landmark group so that we now own it
lmark_group = value.copy()
# check the target is set correctly
lmark_group._group_label = group
else:
raise ValueError('Valid types are PointCloud or LandmarkGroup')
self._landmark_groups[group] = lmark_group
def __getitem__(self, group=None):
"""
Returns the group for the provided label.
Parameters
---------
group : `string`, optional
The label of the group. If None is provided, and if there is only
one group, the unambiguous group will be returned.
Returns
-------
lmark_group : :map:`LandmarkGroup`
The matching landmark group.
"""
if group is None:
if self.n_groups == 1:
group = self.group_labels[0]
else:
raise ValueError("Cannot use None as a key as there are {} "
"landmark groups".format(self.n_groups))
return self._landmark_groups[group]
def __delitem__(self, group):
"""
Delete the group for the provided label.
Parameters
---------
group : `string`
The label of the group.
"""
del self._landmark_groups[group]
def __len__(self):
return len(self._landmark_groups)
@property
def n_groups(self):
"""
Total number of labels.
:type: `int`
"""
return len(self._landmark_groups)
@property
def has_landmarks(self):
"""
Whether the object has landmarks or not
:type: `int`
"""
return self.n_groups != 0
@property
def group_labels(self):
"""
All the labels for the landmark set.
:type: `list` of `str`
"""
return self._landmark_groups.keys()
def _transform_inplace(self, transform):
for group in self._landmark_groups.values():
group.lms._transform_inplace(transform)
return self
def __str__(self):
out_string = '{}: n_groups: {}'.format(type(self).__name__,
self.n_groups)
if self.has_landmarks:
for label in self:
out_string += '\n'
out_string += '({}): {}'.format(label, self[label].__str__())
return out_string
[docs]class LandmarkGroup(MutableMapping, Copyable, Viewable):
r"""
An immutable object that holds a :map:`PointCloud` (or a subclass) and
stores labels for each point. These labels are defined via masks on the
:map:`PointCloud`. For this reason, the :map:`PointCloud` is considered to
be immutable.
The labels to masks must be within an `OrderedDict` so that semantic
ordering can be maintained.
Parameters
----------
pointcloud : :map:`PointCloud`
The pointcloud representing the landmarks.
labels_to_masks : `ordereddict` {`str` -> `bool ndarray`}
For each label, the mask that specifies the indices in to the
pointcloud that belong to the label.
copy : `bool`, optional
If ``True``, a copy of the :map:`PointCloud` is stored on the group.
Raises
------
ValueError
If `dict` passed instead of `OrderedDict`
ValueError
If no set of label masks is passed.
ValueError
If any of the label masks differs in size to the pointcloud.
ValueError
If there exists any point in the pointcloud that is not covered
by a label.
"""
def __init__(self, pointcloud, labels_to_masks, copy=True):
super(LandmarkGroup, self).__init__()
if not labels_to_masks:
raise ValueError('Landmark groups are designed for their internal '
'state, other than owernship, to be immutable. '
'Empty label sets are not permitted.')
if np.vstack(labels_to_masks.values()).shape[1] != pointcloud.n_points:
raise ValueError('Each mask must have the same number of points '
'as the landmark pointcloud.')
if type(labels_to_masks) is dict:
raise ValueError('Must provide an OrderedDict to maintain the '
'semantic meaning of the labels.')
# Another sanity check
self._labels_to_masks = labels_to_masks
self._verify_all_labels_masked()
if copy:
self._pointcloud = pointcloud.copy()
self._labels_to_masks = OrderedDict([(l, m.copy()) for l, m in
labels_to_masks.items()])
else:
self._pointcloud = pointcloud
self._labels_to_masks = labels_to_masks
[docs] def copy(self):
r"""
Generate an efficient copy of this :map:`LandmarkGroup`.
Returns
-------
``type(self)``
A copy of this object
"""
new = Copyable.copy(self)
for k, v in new._labels_to_masks.items():
new._labels_to_masks[k] = v.copy()
return new
def __iter__(self):
"""
Iterate over the internal label dictionary
"""
return iter(self._labels_to_masks)
def __setitem__(self, label, indices):
"""
Add a new label to the landmark group by adding a new set of indices
Parameters
----------
label : `string`
Label of landmark.
indices : ``(K,)`` `ndarray`
Array of indices in to the pointcloud. Each index implies
membership to the label.
"""
mask = np.zeros(self._pointcloud.n_points, dtype=np.bool)
mask[indices] = True
self._labels_to_masks[label] = mask
def __getitem__(self, label=None):
"""
Returns the PointCloud that contains this label represents on the group.
This will be a subset of the total landmark group PointCloud.
Parameters
----------
label : `string`
Label to filter on.
Returns
-------
pcloud : :map:`PointCloud`
The PointCloud that this label represents. Will be a subset of the
entire group's landmarks.
"""
if label is None:
return self.lms.copy()
return self._pointcloud.from_mask(self._labels_to_masks[label])
def __delitem__(self, label):
"""
Delete the semantic labelling for the provided label.
.. note::
You cannot delete a semantic label and leave the landmark group
partially unlabelled. Landmark groups must contain labels for
every point.
Parameters
---------
label : `string`
The label to remove.
Raises
------
ValueError
If deleting the label would leave some points unlabelled
"""
# Pop the value off, which is akin to deleting it (removes it from the
# underlying dict). However, we keep it around so we can check if
# removing it causes an unlabelled point
value_to_delete = self._labels_to_masks.pop(label)
try:
self._verify_all_labels_masked()
except ValueError as e:
# Catch the error, restore the value and re-raise the exception!
self._labels_to_masks[label] = value_to_delete
raise e
def __len__(self):
return len(self._labels_to_masks)
@property
def labels(self):
"""
The list of labels that belong to this group.
:type: `list` of `str`
"""
return self._labels_to_masks.keys()
@property
def n_labels(self):
"""
Number of labels in the group.
:type: `int`
"""
return len(self.labels)
@property
def lms(self):
"""
The pointcloud representing all the landmarks in the group.
:type: :map:`PointCloud`
"""
return self._pointcloud
@property
def n_landmarks(self):
"""
The total number of landmarks in the group.
:type: `int`
"""
return self._pointcloud.n_points
@property
def n_dims(self):
"""
The dimensionality of these landmarks.
:type: `int`
"""
return self._pointcloud.n_dims
[docs] def with_labels(self, labels=None):
"""A new landmark group that contains only the certain labels
Parameters
----------
labels : `str` or `list` of `str`, optional
Labels that should be kept in the returned landmark group. If
``None`` is passed, and if there is only one label on this group,
the label will be substituted automatically.
Returns
-------
landmark_group : :map:`LandmarkGroup`
A new landmark group with the same group label but containing only
the given label.
"""
# make it easier by allowing None when there is only one label
if labels is None:
if self.n_labels == 1:
labels = self.labels
else:
raise ValueError("Cannot use None as there are "
"{} labels".format(self.n_labels))
# Make it easier to use by accepting a single string as well as a list
if isinstance(labels, str):
labels = [labels]
return self._new_group_with_only_labels(labels)
[docs] def without_labels(self, labels):
"""A new landmark group that excludes certain labels
label.
Parameters
----------
labels : `str` or `list` of `str`
Labels that should be excluded in the returned landmark group.
Returns
-------
landmark_group : :map:`LandmarkGroup`
A new landmark group with the same group label but containing all
labels except the given label.
"""
# Make it easier to use by accepting a single string as well as a list
if isinstance(labels, str):
labels = [labels]
labels_to_keep = list(set(self.labels).difference(labels))
return self._new_group_with_only_labels(labels_to_keep)
def _verify_all_labels_masked(self):
"""
Verify that every point in the pointcloud is associated with a label.
If any one point is not covered by a label, then raise a
``ValueError``.
"""
unlabelled_points = np.sum(self._labels_to_masks.values(), axis=0) == 0
if np.any(unlabelled_points):
nonzero = np.nonzero(unlabelled_points)
raise ValueError(
'Every point in the landmark pointcloud must be labelled. '
'Points {0} were unlabelled.'.format(nonzero))
def _new_group_with_only_labels(self, labels):
"""
Deal with changing indices when you add and remove points. In this case
we only deal with building a new dataset that keeps masks.
Parameters
----------
labels : list of `string`
List of strings of the labels to keep
Returns
-------
lmark_group : :map:`LandmarkGroup`
The new landmark group with only the requested labels.
"""
set_difference = set(labels).difference(self.labels)
if len(set_difference) > 0:
raise ValueError('Labels {0} do not exist in the landmark '
'group. Available labels are: {1}'.format(
list(set_difference), self.labels))
masks_to_keep = [self._labels_to_masks[l] for l in labels
if l in self._labels_to_masks]
overlap = np.sum(masks_to_keep, axis=0) > 0
masks_to_keep = [l[overlap] for l in masks_to_keep]
return LandmarkGroup(self._pointcloud.from_mask(overlap),
OrderedDict(zip(labels, masks_to_keep)))
[docs] def tojson(self):
r"""
Convert this `LandmarkGroup` to a dictionary JSON representation.
Returns
-------
json : ``dict``
Dictionary conforming to the LJSON v2 specification.
"""
labels = [{'mask': mask.nonzero()[0].tolist(),
'label': label}
for label, mask in self._labels_to_masks.items()]
return {'landmarks': self.lms.tojson(),
'labels': labels}
def _view_2d(self, with_labels=None, without_labels=None, group='group',
figure_id=None, new_figure=False, image_view=True,
render_lines=True, line_colour=None, line_style='-',
line_width=1, render_markers=True, marker_style='o',
marker_size=20, marker_face_colour=None,
marker_edge_colour=None, marker_edge_width=1.,
render_numbering=False, numbers_horizontal_align='center',
numbers_vertical_align='bottom',
numbers_font_name='sans-serif', numbers_font_size=10,
numbers_font_style='normal', numbers_font_weight='normal',
numbers_font_colour='k', render_legend=True, legend_title='',
legend_font_name='sans-serif', legend_font_style='normal',
legend_font_size=10, legend_font_weight='normal',
legend_marker_scale=None, legend_location=2,
legend_bbox_to_anchor=(1.05, 1.), legend_border_axes_pad=None,
legend_n_columns=1, legend_horizontal_spacing=None,
legend_vertical_spacing=None, legend_border=True,
legend_border_padding=None, legend_shadow=False,
legend_rounded_corners=False, render_axes=True,
axes_font_name='sans-serif', axes_font_size=10,
axes_font_style='normal', axes_font_weight='normal',
axes_x_limits=None, axes_y_limits=None, figure_size=(10, 8)):
"""
Visualize the landmark group.
Parameters
----------
with_labels : ``None`` or `str` or `list` of `str`, optional
If not ``None``, only show the given label(s). Should **not** be
used with the ``without_labels`` kwarg.
without_labels : ``None`` or `str` or `list` of `str`, optional
If not ``None``, show all except the given label(s). Should **not**
be used with the ``with_labels`` kwarg.
group : `str` or `None`, optional
The landmark group to be visualized. If ``None`` and there are more
than one landmark groups, an error is raised.
figure_id : `object`, optional
The id of the figure to be used.
new_figure : `bool`, optional
If ``True``, a new figure is created.
image_view : `bool`, optional
If ``True``, the x and y axes are flipped.
render_lines : `bool`, optional
If ``True``, the edges will be rendered.
line_colour : {``r``, ``g``, ``b``, ``c``, ``m``, ``k``, ``w``} or
``(3, )`` `ndarray` or ``None``, optional
The colour of the lines. If ``None``, a different colour will be
automatically selected for each label.
line_style : {``-``, ``--``, ``-.``, ``:``}, optional
The style of the lines.
line_width : `float`, optional
The width of the lines.
render_markers : `bool`, optional
If ``True``, the markers will be rendered.
marker_style : {``.``, ``,``, ``o``, ``v``, ``^``, ``<``, ``>``, ``+``,
``x``, ``D``, ``d``, ``s``, ``p``, ``*``, ``h``, ``H``,
``1``, ``2``, ``3``, ``4``, ``8``}, optional
The style of the markers.
marker_size : `int`, optional
The size of the markers in points^2.
marker_face_colour : {``r``, ``g``, ``b``, ``c``, ``m``, ``k``, ``w``}
or ``(3, )`` `ndarray`, optional
The face (filling) colour of the markers.
marker_edge_colour : {``r``, ``g``, ``b``, ``c``, ``m``, ``k``, ``w``}
or ``(3, )`` `ndarray`, optional
The edge colour of the markers.
marker_edge_width : `float`, optional
The width of the markers' edge.
render_numbering : `bool`, optional
If ``True``, the landmarks will be numbered.
numbers_horizontal_align : {``center``, ``right``, ``left``}, optional
The horizontal alignment of the numbers' texts.
numbers_vertical_align : {``center``, ``top``, ``bottom``,
``baseline``}, optional
The vertical alignment of the numbers' texts.
numbers_font_name : {``serif``, ``sans-serif``, ``cursive``,
``fantasy``, ``monospace``}, optional
The font of the numbers.
numbers_font_size : `int`, optional
The font size of the numbers.
numbers_font_style : {``normal``, ``italic``, ``oblique``}, optional
The font style of the numbers.
numbers_font_weight : {``ultralight``, ``light``, ``normal``,
``regular``, ``book``, ``medium``, ``roman``,
``semibold``, ``demibold``, ``demi``, ``bold``,
``heavy``, ``extra bold``, ``black``}, optional
The font weight of the numbers.
numbers_font_colour : {``r``, ``g``, ``b``, ``c``, ``m``, ``k``, ``w``}
or ``(3, )`` `ndarray`, optional
The font colour of the numbers.
render_legend : `bool`, optional
If ``True``, the legend will be rendered.
legend_title : `str`, optional
The title of the legend.
legend_font_name : {``serif``, ``sans-serif``, ``cursive``,
``fantasy``, ``monospace``}, optional
The font of the legend.
legend_font_style : {``normal``, ``italic``, ``oblique``}, optional
The font style of the legend.
legend_font_size : `int`, optional
The font size of the legend.
legend_font_weight : {``ultralight``, ``light``, ``normal``,
``regular``, ``book``, ``medium``, ``roman``,
``semibold``, ``demibold``, ``demi``, ``bold``,
``heavy``, ``extra bold``, ``black``}, optional
The font weight of the legend.
legend_marker_scale : `float`, optional
The relative size of the legend markers with respect to the original
legend_location : `int`, optional
The location of the legend. The predefined values are:
=============== ===
'best' 0
'upper right' 1
'upper left' 2
'lower left' 3
'lower right' 4
'right' 5
'center left' 6
'center right' 7
'lower center' 8
'upper center' 9
'center' 10
=============== ===
legend_bbox_to_anchor : (`float`, `float`), optional
The bbox that the legend will be anchored.
legend_border_axes_pad : `float`, optional
The pad between the axes and legend border.
legend_n_columns : `int`, optional
The number of the legend's columns.
legend_horizontal_spacing : `float`, optional
The spacing between the columns.
legend_vertical_spacing : `float`, optional
The vertical space between the legend entries.
legend_border : `bool`, optional
If ``True``, a frame will be drawn around the legend.
legend_border_padding : `float`, optional
The fractional whitespace inside the legend border.
legend_shadow : `bool`, optional
If ``True``, a shadow will be drawn behind legend.
legend_rounded_corners : `bool`, optional
If ``True``, the frame's corners will be rounded (fancybox).
render_axes : `bool`, optional
If ``True``, the axes will be rendered.
axes_font_name : {``serif``, ``sans-serif``, ``cursive``, ``fantasy``,
``monospace``}, optional
The font of the axes.
axes_font_size : `int`, optional
The font size of the axes.
axes_font_style : {``normal``, ``italic``, ``oblique``}, optional
The font style of the axes.
axes_font_weight : {``ultralight``, ``light``, ``normal``, ``regular``,
``book``, ``medium``, ``roman``, ``semibold``,
``demibold``, ``demi``, ``bold``, ``heavy``,
``extra bold``, ``black``}, optional
The font weight of the axes.
axes_x_limits : (`float`, `float`) or `None`, optional
The limits of the x axis.
axes_y_limits : (`float`, `float`) or `None`, optional
The limits of the y axis.
figure_size : (`float`, `float`) or `None`, optional
The size of the figure in inches.
Raises
------
ValueError
If both ``with_labels`` and ``without_labels`` are passed.
"""
from menpo.visualize import LandmarkViewer2d
if with_labels is not None and without_labels is not None:
raise ValueError('You may only pass one of `with_labels` or '
'`without_labels`.')
elif with_labels is not None:
lmark_group = self.with_labels(with_labels)
elif without_labels is not None:
lmark_group = self.without_labels(without_labels)
else:
lmark_group = self # Fall through
landmark_viewer = LandmarkViewer2d(figure_id, new_figure,
group, lmark_group._pointcloud,
lmark_group._labels_to_masks)
return landmark_viewer.render(
image_view=image_view, render_lines=render_lines,
line_colour=line_colour, line_style=line_style,
line_width=line_width, render_markers=render_markers,
marker_style=marker_style, marker_size=marker_size,
marker_face_colour=marker_face_colour,
marker_edge_colour=marker_edge_colour,
marker_edge_width=marker_edge_width,
render_numbering=render_numbering,
numbers_horizontal_align=numbers_horizontal_align,
numbers_vertical_align=numbers_vertical_align,
numbers_font_name=numbers_font_name,
numbers_font_size=numbers_font_size,
numbers_font_style=numbers_font_style,
numbers_font_weight=numbers_font_weight,
numbers_font_colour=numbers_font_colour,
render_legend=render_legend, legend_title=legend_title,
legend_font_name=legend_font_name,
legend_font_style=legend_font_style,
legend_font_size=legend_font_size,
legend_font_weight=legend_font_weight,
legend_marker_scale=legend_marker_scale,
legend_location=legend_location,
legend_bbox_to_anchor=legend_bbox_to_anchor,
legend_border_axes_pad=legend_border_axes_pad,
legend_n_columns=legend_n_columns,
legend_horizontal_spacing=legend_horizontal_spacing,
legend_vertical_spacing=legend_vertical_spacing,
legend_border=legend_border,
legend_border_padding=legend_border_padding,
legend_shadow=legend_shadow,
legend_rounded_corners=legend_rounded_corners,
render_axes=render_axes, axes_font_name=axes_font_name,
axes_font_size=axes_font_size, axes_font_style=axes_font_style,
axes_font_weight=axes_font_weight, axes_x_limits=axes_x_limits,
axes_y_limits=axes_y_limits, figure_size=figure_size)
def _view_3d(self, figure_id=None, new_figure=False, **kwargs):
try:
from menpo3d.visualize import LandmarkViewer3d
return LandmarkViewer3d(figure_id, new_figure,
self._pointcloud, self).render(**kwargs)
except ImportError:
from menpo.visualize import Menpo3dErrorMessage
raise ImportError(Menpo3dErrorMessage)
def __str__(self):
return '{}: n_labels: {}, n_points: {}'.format(
type(self).__name__, self.n_labels, self.n_landmarks)