Class StdShiftIntensity

monai/transforms/intensity/array.py:330–381

Shift intensity for the image with a factor and the standard deviation of the image by: ``v = v + factor * std(v)``. This transform can focus on only non-zero values or the entire image, and can also calculate the std on each channel separately. Args: factor: factor shi

class StdShiftIntensity(Transform):
    """
    Shift intensity for the image with a factor and the standard deviation of the image
    by: ``v = v + factor * std(v)``.
    This transform can focus on only non-zero values or the entire image,
    and can also calculate the std on each channel separately.

    Args:
        factor: factor shift by ``v = v + factor * std(v)``.
        nonzero: whether only count non-zero values.
        channel_wise: if True, calculate on each channel separately. Please ensure
            that the first dimension represents the channel of the image if True.
        dtype: output data type, if None, same as input image. defaults to float32.
    """

    backend = [TransformBackends.TORCH, TransformBackends.NUMPY]

    def __init__(
        self, factor: float, nonzero: bool = False, channel_wise: bool = False, dtype: DtypeLike = np.float32
    ) -> None:
        self.factor = factor
        self.nonzero = nonzero
        self.channel_wise = channel_wise
        self.dtype = dtype

    def _stdshift(self, img: NdarrayOrTensor) -> NdarrayOrTensor:
        ones: Callable
        std: Callable
        if isinstance(img, torch.Tensor):
            ones = torch.ones
            std = partial(torch.std, unbiased=False)
        else:
            ones = np.ones
            std = np.std

        slices = (img != 0) if self.nonzero else ones(img.shape, dtype=bool)
        if slices.any():
            offset = self.factor * std(img[slices])
            img[slices] = img[slices] + offset
        return img

    def __call__(self, img: NdarrayOrTensor) -> NdarrayOrTensor:
        """
        Apply the transform to `img`.
        """
        img = convert_to_tensor(img, track_meta=get_track_meta(), dtype=self.dtype)
        if self.channel_wise:
            for i, d in enumerate(img):
                img[i] = self._stdshift(d)  # type: ignore
        else:
            img = self._stdshift(img)
        return img


class RandStdShiftIntensity(RandomizableTransform):