Function rotate90

monai/transforms/spatial/functional.py:540–587

Functional implementation of rotate90. This function operates eagerly or lazily according to ``lazy`` (default ``False``). Args: img: data to be changed, assuming `img` is channel-first. axes: 2 int numbers, defines the plane to rotate with 2 spatial axes.

(img, axes, k, lazy, transform_info)

def rotate90(img, axes, k, lazy, transform_info):
    """
    Functional implementation of rotate90.
    This function operates eagerly or lazily according to
    ``lazy`` (default ``False``).

    Args:
        img: data to be changed, assuming `img` is channel-first.
        axes: 2 int numbers, defines the plane to rotate with 2 spatial axes.
            If axis is negative it counts from the last to the first axis.
        k: number of times to rotate by 90 degrees.
        lazy: a flag that indicates whether the operation should be performed lazily or not
        transform_info: a dictionary with the relevant information pertaining to an applied transform.
    """
    extra_info = {"axes": [d - 1 for d in axes], "k": k}
    ori_shape = img.peek_pending_shape() if isinstance(img, MetaTensor) else img.shape[1:]
    sp_shape = list(ori_shape)
    if k in (1, 3):
        a_0, a_1 = axes[0] - 1, axes[1] - 1
        sp_shape[a_0], sp_shape[a_1] = ori_shape[a_1], ori_shape[a_0]
    rank = img.peek_pending_rank() if isinstance(img, MetaTensor) else torch.tensor(3.0, dtype=torch.double)
    r, sp_r = int(rank), len(ori_shape)
    xform = to_affine_nd(r, create_translate(sp_r, [-float(d - 1) / 2 for d in sp_shape]))
    s = -1.0 if int(axes[0]) - int(axes[1]) in (-1, 2) else 1.0
    if sp_r == 2:
        rot90 = to_affine_nd(r, create_rotate(sp_r, [s * np.pi / 2]))
    else:
        idx = {1, 2, 3} - set(axes)
        angle: list[float] = [0, 0, 0]
        angle[idx.pop() - 1] = s * np.pi / 2
        rot90 = to_affine_nd(r, create_rotate(sp_r, angle))
    for _ in range(k):
        xform = rot90 @ xform
    xform = to_affine_nd(r, create_translate(sp_r, [float(d - 1) / 2 for d in ori_shape])) @ xform
    meta_info = TraceableTransform.track_transform_meta(
        img,
        sp_size=sp_shape,
        affine=xform,
        extra_info=extra_info,
        orig_size=ori_shape,
        transform_info=transform_info,
        lazy=lazy,
    )
    out = _maybe_new_metatensor(img)
    if lazy:
        return out.copy_meta_from(meta_info) if isinstance(out, MetaTensor) else meta_info
    out = torch.rot90(out, k, axes)
    return out.copy_meta_from(meta_info) if isinstance(out, MetaTensor) else out


def affine_func(