Class GaussianSharpen

monai/transforms/intensity/array.py:1703–1756

Sharpen images using the Gaussian Blur filter. Referring to: http://scipy-lectures.org/advanced/image_processing/auto_examples/plot_sharpen.html. The algorithm is shown as below .. code-block:: python blurred_f = gaussian_filter(img, sigma1) filter_blurred_f = gaus

class GaussianSharpen(Transform):
    """
    Sharpen images using the Gaussian Blur filter.
    Referring to: http://scipy-lectures.org/advanced/image_processing/auto_examples/plot_sharpen.html.
    The algorithm is shown as below

    .. code-block:: python

        blurred_f = gaussian_filter(img, sigma1)
        filter_blurred_f = gaussian_filter(blurred_f, sigma2)
        img = blurred_f + alpha * (blurred_f - filter_blurred_f)

    A set of default values `sigma1=3.0`, `sigma2=1.0` and `alpha=30.0` is provide for reference.

    Args:
        sigma1: sigma parameter for the first gaussian kernel. 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.
        sigma2: sigma parameter for the second gaussian kernel. 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.
        alpha: weight parameter to compute the final result.
        approx: discrete Gaussian kernel type, available options are "erf", "sampled", and "scalespace".
            see also :py:meth:`monai.networks.layers.GaussianFilter`.

    """

    backend = [TransformBackends.TORCH]

    def __init__(
        self,
        sigma1: Sequence[float] | float = 3.0,
        sigma2: Sequence[float] | float = 1.0,
        alpha: float = 30.0,
        approx: str = "erf",
    ) -> None:
        self.sigma1 = sigma1
        self.sigma2 = sigma2
        self.alpha = alpha
        self.approx = approx

    def __call__(self, img: NdarrayTensor) -> NdarrayTensor:
        img = convert_to_tensor(img, track_meta=get_track_meta())
        img_t, *_ = convert_data_type(img, torch.Tensor, dtype=torch.float32)

        gf1, gf2 = (
            GaussianFilter(img_t.ndim - 1, sigma, approx=self.approx).to(img_t.device)
            for sigma in (self.sigma1, self.sigma2)
        )
        blurred_f = gf1(img_t.unsqueeze(0))
        filter_blurred_f = gf2(blurred_f)
        out_t: torch.Tensor = (blurred_f + self.alpha * (blurred_f - filter_blurred_f)).squeeze(0)
        out, *_ = convert_to_dst_type(out_t, dst=img, dtype=out_t.dtype)
        return out


class RandGaussianSharpen(RandomizableTransform):