Method __call__

monai/transforms/croppad/array.py:995–1038

Args: img: input image to sample patches from. assuming `img` is a channel-first array. weight_map: weight map used to generate patch samples. The weights must be non-negative. Each element denotes a sampling weight of the spatial location. 0 indicate

(
        self,
        img: torch.Tensor,
        weight_map: NdarrayOrTensor | None = None,
        randomize: bool = True,
        lazy: bool | None = None,
    )

        self._lazy = _val

    def __call__(
        self,
        img: torch.Tensor,
        weight_map: NdarrayOrTensor | None = None,
        randomize: bool = True,
        lazy: bool | None = None,
    ) -> list[torch.Tensor]:
        """
        Args:
            img: input image to sample patches from. assuming `img` is a channel-first array.
            weight_map: weight map used to generate patch samples. The weights must be non-negative.
                Each element denotes a sampling weight of the spatial location. 0 indicates no sampling.
                It should be a single-channel array in shape, for example, `(1, spatial_dim_0, spatial_dim_1, ...)`
            randomize: whether to execute random operations, default to `True`.
            lazy: a flag to override the lazy behaviour for this call, if set. Defaults to None.

        Returns:
            A list of image patches
        """
        img_shape = img.peek_pending_shape() if isinstance(img, MetaTensor) else img.shape[1:]

        if randomize:
            if weight_map is None:
                weight_map = self.weight_map
            if weight_map is None:
                raise ValueError("weight map must be provided for weighted patch sampling.")
            w_shape = weight_map.peek_pending_shape() if isinstance(weight_map, MetaTensor) else weight_map.shape[1:]
            if img_shape != w_shape:
                warnings.warn(f"image and weight map spatial shape mismatch: {img_shape} vs {w_shape}.")
            self.randomize(weight_map)

        _spatial_size = fall_back_tuple(self.spatial_size, img_shape)
        results: list[torch.Tensor] = []
        lazy_ = self.lazy if lazy is None else lazy
        for i, center in enumerate(self.centers):
            cropper = SpatialCrop(roi_center=center, roi_size=_spatial_size, lazy=lazy_)
            cropped = cropper(img)
            if get_track_meta():
                ret_: MetaTensor = cropped  # type: ignore
                ret_.meta[Key.PATCH_INDEX] = i
                ret_.meta["crop_center"] = center
                self.push_transform(ret_, replace=True, lazy=lazy_)
            results.append(cropped)
        return results


class RandCropByPosNegLabel(Randomizable, TraceableTransform, LazyTransform, MultiSampleTrait):