Class AsDiscrete

monai/transforms/post/array.py:140–244

Convert the input tensor/array into discrete values, possible operations are: - `argmax`. - threshold input value to binary values. - convert input value to One-Hot format (set ``to_one_hot=N``, `N` is the number of classes). - round the value to the closest

class AsDiscrete(Transform):
    """
    Convert the input tensor/array into discrete values, possible operations are:

        -  `argmax`.
        -  threshold input value to binary values.
        -  convert input value to One-Hot format (set ``to_one_hot=N``, `N` is the number of classes).
        -  round the value to the closest integer.

    Args:
        argmax: whether to execute argmax function on input data before transform.
            Defaults to ``False``.
        to_onehot: if not None, convert input data into the one-hot format with specified number of classes.
            Defaults to ``None``.
        threshold: if not None, threshold the float values to int number 0 or 1 with specified threshold.
            Defaults to ``None``.
        rounding: if not None, round the data according to the specified option,
            available options: ["torchrounding"].
        kwargs: additional parameters to `torch.argmax`, `monai.networks.one_hot`.
            currently ``dim``, ``keepdim``, ``dtype`` are supported, unrecognized parameters will be ignored.
            These default to ``0``, ``True``, ``torch.float`` respectively.

    Example:

        >>> transform = AsDiscrete(argmax=True)
        >>> print(transform(np.array([[[0.0, 1.0]], [[2.0, 3.0]]])))
        # [[[1.0, 1.0]]]

        >>> transform = AsDiscrete(threshold=0.6)
        >>> print(transform(np.array([[[0.0, 0.5], [0.8, 3.0]]])))
        # [[[0.0, 0.0], [1.0, 1.0]]]

        >>> transform = AsDiscrete(argmax=True, to_onehot=2, threshold=0.5)
        >>> print(transform(np.array([[[0.0, 1.0]], [[2.0, 3.0]]])))
        # [[[0.0, 0.0]], [[1.0, 1.0]]]

    """

    backend = [TransformBackends.TORCH]

    def __init__(
        self,
        argmax: bool = False,
        to_onehot: int | None = None,
        threshold: float | None = None,
        rounding: str | None = None,
        **kwargs,
    ) -> None:
        self.argmax = argmax
        if isinstance(to_onehot, bool):  # for backward compatibility
            raise ValueError("`to_onehot=True/False` is deprecated, please use `to_onehot=num_classes` instead.")
        self.to_onehot = to_onehot
        self.threshold = threshold
        self.rounding = rounding
        self.kwargs = kwargs

    def __call__(
        self,