Transform¶

class menpo.transform.Transform[source]

Bases: Copyable

Abstract representation of any spatial transform.

Provides a unified interface to apply the transform with apply_inplace() and apply().

All Transforms support basic composition to form a TransformChain.

There are two useful forms of composition. Firstly, the mathematical composition symbol o has the following definition:

Let a(x) and b(x) be two transforms on x.
(a o b)(x) == a(b(x))


This functionality is provided by the compose_after() family of methods:

(a.compose_after(b)).apply(x) == a.apply(b.apply(x))


Equally useful is an inversion the order of composition - so that over time a large chain of transforms can be built to do a useful job, and composing on this chain adds another transform to the end (after all other preceding transforms have been performed).

For instance, let’s say we want to rescale a PointCloud p around its mean, and then translate it some place else. It would be nice to be able to do something like:

t = Translation(-p.centre)  # translate to centre
s = Scale(2.0)  # rescale
move = Translate([10, 0 ,0])  # budge along the x axis
t.compose(s).compose(-t).compose(move)


In Menpo, this functionality is provided by the compose_before() family of methods:

(a.compose_before(b)).apply(x) == b.apply(a.apply(x))


For native composition, see the ComposableTransform subclass and the VComposable mix-in.

For inversion, see the Invertible and VInvertible mix-ins.

For alignment, see the Alignment mix-in.

apply(x, batch_size=None, **kwargs)[source]

Applies this transform to x.

If x is Transformable, x will be handed this transform object to transform itself non-destructively (a transformed copy of the object will be returned).

If not, x is assumed to be an ndarray. The transformation will be non-destructive, returning the transformed version.

Any kwargs will be passed to the specific transform _apply() method.

Parameters: x (Transformable or (n_points, n_dims) ndarray) – The array or object to be transformed. batch_size (int, optional) – If not None, this determines how many items from the numpy array will be passed through the transform at a time. This is useful for operations that require large intermediate matrices to be computed. kwargs (dict) – Passed through to _apply(). transformed (type(x)) – The transformed object or array
apply_inplace(*args, **kwargs)[source]

Deprecated as public supported API, use the non-mutating apply() instead.

For internal performance-specific uses, see _apply_inplace().

compose_after(transform)[source]

Returns a TransformChain that represents this transform composed after the given transform:

c = a.compose_after(b)
c.apply(p) == a.apply(b.apply(p))


a and b are left unchanged.

This corresponds to the usual mathematical formalism for the compose operator, o.

Parameters: transform (Transform) – Transform to be applied before self transform (TransformChain) – The resulting transform chain.
compose_before(transform)[source]

Returns a TransformChain that represents this transform composed before the given transform:

c = a.compose_before(b)
c.apply(p) == b.apply(a.apply(p))


a and b are left unchanged.

Parameters: transform (Transform) – Transform to be applied after self transform (TransformChain) – The resulting transform chain.
copy()

Generate an efficient copy of this object.

Note that Numpy arrays and other Copyable objects on self will be deeply copied. Dictionaries and sets will be shallow copied, and everything else will be assigned (no copy will be made).

Classes that store state other than numpy arrays and immutable types should overwrite this method to ensure all state is copied.

Returns: type(self) – A copy of this object
n_dims

The dimensionality of the data the transform operates on.

None if the transform is not dimension specific.

Type: int or None
n_dims_output

The output of the data from the transform.

None if the output of the transform is not dimension specific.

Type: int or None