Skip to content

OneHot

Bases: Transform

One-hot encode label maps.

Each label map with \(K\) classes (including background) is converted from shape \((1, I, J, K)\) to \((K, I, J, K)\), where channel \(k\) is 1 where the label equals \(k\) and 0 elsewhere.

Only LabelMap images are affected. ScalarImage instances are left unchanged.

The inverse takes the argmax across channels, restoring the original single-channel label map.

Parameters:

Name Type Description Default
num_classes int

Total number of classes. -1 (default) infers from the data as max_label + 1.

 -1
**kwargs Any

See Transform.

 {}

Examples:

>>> import torchio as tio
>>> transform = tio.OneHot()
>>> transform = tio.OneHot(num_classes=5)
>>> # Invert back to single-channel
>>> restored = transformed.apply_inverse_transform()
Source code in src/torchio/transforms/label/one_hot.py 
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
class OneHot(Transform):
    r"""One-hot encode label maps.

    Each label map with $K$ classes (including background) is converted
    from shape $(1, I, J, K)$ to $(K, I, J, K)$, where channel $k$
    is 1 where the label equals $k$ and 0 elsewhere.

    Only [`LabelMap`][torchio.LabelMap] images are affected.
    [`ScalarImage`][torchio.ScalarImage] instances are left unchanged.

    The inverse takes the argmax across channels, restoring the
    original single-channel label map.

    Args:
        num_classes: Total number of classes. `-1` (default) infers
            from the data as `max_label + 1`.
        **kwargs: See [`Transform`][torchio.Transform].

    Examples:
        >>> import torchio as tio
        >>> transform = tio.OneHot()
        >>> transform = tio.OneHot(num_classes=5)
        >>> # Invert back to single-channel
        >>> restored = transformed.apply_inverse_transform()
    """

    def __init__(
        self,
        *,
        num_classes: int = -1,
        **kwargs: Any,
    ) -> None:
        super().__init__(**kwargs)
        self.num_classes = num_classes

    def make_params(self, batch: SubjectsBatch) -> dict[str, Any]:
        """No random parameters."""
        return {"num_classes": self.num_classes}

    def apply_transform(
        self,
        batch: SubjectsBatch,
        params: dict[str, Any],
    ) -> SubjectsBatch:
        """One-hot encode each label map in the batch."""
        num_classes = params["num_classes"]
        for _name, img_batch in batch.images.items():
            if not issubclass(img_batch._image_class, LabelMap):
                continue
            # (B, 1, I, J, K) -> (B, num_classes, I, J, K)
            data = img_batch.data.long()
            flat = data[:, 0]  # (B, I, J, K)
            encoded = functional.one_hot(flat, num_classes=num_classes)
            # (B, I, J, K, num_classes) -> (B, num_classes, I, J, K)
            img_batch.data = encoded.permute(0, 4, 1, 2, 3).float()
        return batch

    @property
    def invertible(self) -> bool:
        """Whether this transform can be inverted."""
        return True

    def inverse(self, params: dict[str, Any]) -> _OneHotInverse:
        """Invert by taking argmax."""
        return _OneHotInverse(copy=False)

supports_per_instance_params property

Whether this transform can sample parameters per batch element.

Defaults to False. Transforms that implement per-instance parameter sampling override this to return True. When False, the transform always uses batch-shared parameters regardless of the per_instance flag, preserving the legacy behavior.

supports_per_instance_p property

Whether this transform can gate each batch element independently.

Defaults to False. Shape-preserving transforms that implement per-element probability override this to return True. Shape-changing transforms must leave it False because masked and unmasked elements would have incompatible shapes.

invertible property

Whether this transform can be inverted.

forward(data) 

forward(data: Subject) -> Subject

forward(data: Image) -> Image

forward(data: Tensor) -> Tensor

forward(data: np.ndarray) -> np.ndarray

forward(data: sitk.Image) -> sitk.Image

forward(data: nib.Nifti1Image) -> nib.Nifti1Image

forward(data: dict) -> dict

forward(data: ImagesBatch) -> ImagesBatch

forward(data: SubjectsBatch) -> SubjectsBatch

Apply the transform.

The output type always matches the input type.

Parameters:

Name Type Description Default
data Any

Input data to transform.

 required
Source code in src/torchio/transforms/transform.py 
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
def forward(self, data: Any) -> Any:
    """Apply the transform.

    The output type always matches the input type.

    Args:
        data: Input data to transform.
    """
    if self.copy:
        data = _copy.deepcopy(data)
    batch, unwrap = self._wrap(data)
    # When per-element gating is active, the transform handles the
    # probability itself (masked-out elements get identity params),
    # so skip the batch-wide coin flip here. Apply iff rand < p, so
    # p=0 is always a no-op and p=1 always applies.
    if not self._per_instance_p_active(batch) and torch.rand(1).item() >= self.p:
        return unwrap(batch)
    params = self.make_params(batch)
    batch = self.apply_transform(batch, params)
    # Record history on the batch, unless every element was gated out by
    # per-element probability: that is an exact no-op, and recording it
    # would let history replay (e.g. an invertible spatial transform)
    # trigger an unnecessary identity resample.
    if not _all_elements_gated_out(params):
        trace = AppliedTransform(name=type(self).__name__, params=params)
        if not hasattr(batch, "applied_transforms"):
            batch.applied_transforms = []
        batch.applied_transforms.append(trace)
    result = unwrap(batch)
    # Propagate history to outputs that can carry it
    if (
        hasattr(batch, "applied_transforms")
        and not isinstance(result, (SubjectsBatch, Tensor, np.ndarray))
        and not isinstance(result, dict)
    ):
        with contextlib.suppress(AttributeError):
            result.applied_transforms = list(batch.applied_transforms)
    return result

to_hydra()

Export as a Hydra-compatible config dict.

Returns a dict with _target_ set to the fully qualified class name and only non-default field values included.

Returns:

Type Description
dict[str, Any]

Dict suitable for hydra.utils.instantiate().
Source code in src/torchio/transforms/transform.py 
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
def to_hydra(self) -> dict[str, Any]:
    """Export as a Hydra-compatible config dict.

    Returns a dict with `_target_` set to the fully qualified
    class name and only non-default field values included.

    Returns:
        Dict suitable for `hydra.utils.instantiate()`.
    """
    from .parameter_range import _ParameterRange

    cls = type(self)
    target = f"torchio.{cls.__qualname__}"
    cfg: dict[str, Any] = {"_target_": target}

    for name, default in _collect_init_params(cls).items():
        value = getattr(self, name, default)
        if isinstance(value, _ParameterRange):
            if value._original == default:
                continue
            value = _hydra_value(value._original)
        elif value == default:
            continue
        else:
            value = _hydra_value(value)
        cfg[name] = value
    return cfg

make_params(batch)

No random parameters.

Source code in src/torchio/transforms/label/one_hot.py 
49
50
51
def make_params(self, batch: SubjectsBatch) -> dict[str, Any]:
    """No random parameters."""
    return {"num_classes": self.num_classes}

apply_transform(batch, params)

One-hot encode each label map in the batch.

Source code in src/torchio/transforms/label/one_hot.py 
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
def apply_transform(
    self,
    batch: SubjectsBatch,
    params: dict[str, Any],
) -> SubjectsBatch:
    """One-hot encode each label map in the batch."""
    num_classes = params["num_classes"]
    for _name, img_batch in batch.images.items():
        if not issubclass(img_batch._image_class, LabelMap):
            continue
        # (B, 1, I, J, K) -> (B, num_classes, I, J, K)
        data = img_batch.data.long()
        flat = data[:, 0]  # (B, I, J, K)
        encoded = functional.one_hot(flat, num_classes=num_classes)
        # (B, I, J, K, num_classes) -> (B, num_classes, I, J, K)
        img_batch.data = encoded.permute(0, 4, 1, 2, 3).float()
    return batch

inverse(params)

Invert by taking argmax.

Source code in src/torchio/transforms/label/one_hot.py 
76
77
78
def inverse(self, params: dict[str, Any]) -> _OneHotInverse:
    """Invert by taking argmax."""
    return _OneHotInverse(copy=False)