e input image. If unset, will use the inferred format from the input.

    Returns:
        `tuple`: The target (height, width) dimension of the output image after resizing.
    r