def __call__(
self,
img: np.ndarray,
label: Optional[np.ndarray] = None,
sigma: Union[Sequence[float], float, Sequence[torch.Tensor],
torch.Tensor] = 3.0,
rescale_min: float = -1.0,
rescale_max: float = 1.0,
):
"""
Args:
img: the image that we want to add new channel to.
label: label image to get extreme points from. Shape must be
(1, spatial_dim1, [, spatial_dim2, ...]). Doesn't support one-hot labels.
sigma: if a list of values, must match the count of spatial dimensions of input data,
and apply every value in the list to 1 spatial dimension. if only 1 value provided,
use it for all spatial dimensions.
rescale_min: minimum value of output data.
rescale_max: maximum value of output data.
"""
if label is None:
raise ValueError("This transform requires a label array!")
if label.shape[0] != 1:
raise ValueError("Only supports single channel labels!")
# Generate extreme points
self.randomize(label[0, :])
points_image = extreme_points_to_image(points=self._points,
label=label,
sigma=sigma,
rescale_min=rescale_min,
rescale_max=rescale_max)
return np.concatenate([img, points_image], axis=0)
def __call__(self, data):
d = dict(data)
label = d[self.label_key]
if label.shape[0] != 1:
raise ValueError("Only supports single channel labels!")
# Generate extreme points
self.randomize(label[0, :])
for key in self.key_iterator(d):
img = d[key]
points_image = extreme_points_to_image(
points=self.points,
label=label,
sigma=self.sigma,
rescale_min=self.rescale_min,
rescale_max=self.rescale_max,
)
d[key] = np.concatenate([img, points_image], axis=0)
return d
