Function rotmat_to_rot6d

detrsmpl/utils/transforms.py:165–184

Convert rotation matrixs to rotation 6d representations. Args: matrix (Union[torch.Tensor, numpy.ndarray]): input shape should be (..., 3, 3). ndim of input is unlimited. Returns: Union[torch.Tensor, numpy.ndarray]: shape would be (..., 6). [1] Zhou, Y.,

(
    matrix: Union[torch.Tensor, numpy.ndarray]
)

def rotmat_to_rot6d(
    matrix: Union[torch.Tensor, numpy.ndarray]
) -> Union[torch.Tensor, numpy.ndarray]:
    """Convert rotation matrixs to rotation 6d representations.

    Args:
        matrix (Union[torch.Tensor, numpy.ndarray]): input shape
                should be (..., 3, 3). ndim of input is unlimited.
    Returns:
        Union[torch.Tensor, numpy.ndarray]: shape would be (..., 6).

    [1] Zhou, Y., Barnes, C., Lu, J., Yang, J., & Li, H.
    On the Continuity of Rotation Representations in Neural Networks.
    IEEE Conference on Computer Vision and Pattern Recognition, 2019.
    Retrieved from http://arxiv.org/abs/1812.07035
    """
    if matrix.shape[-1] != 3 or matrix.shape[-2] != 3:
        raise ValueError(f'Invalid rotation matrix  shape f{matrix.shape}.')
    t = Compose([matrix_to_rotation_6d])
    return t(matrix)


def quat_to_aa(