# symmetric_keypoints_flip Python API: `luxonis_ml.data.augmentations.custom.symmetric_keypoints_flip` ## Classes ### HorizontalSymmetricKeypointsFlip Flip images and symmetric keypoints horizontally. #### Methods ##### init ```python def __init__(keypoint_pairs: list[tuple[int, int]], p: float = 0.5): ``` Flip an image and symmetric keypoints horizontally. Bounding boxes and segmentation masks are flipped as well. Parameters * `keypoint_pairs` (`list[tuple[int, int]]`): Pairs of keypoint indices to swap after the flip. * `p` (`float`): Probability of applying the augmentation. ##### apply ```python def apply(img: np.ndarray, **params) -> np.ndarray: ``` Flip an image horizontally. Parameters * `img` (`np.ndarray`): Image to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped image. ##### apply_to_bboxes ```python def apply_to_bboxes(bboxes: np.ndarray, **params) -> np.ndarray: ``` Flip bounding boxes horizontally. Parameters * `bboxes` (`np.ndarray`): Bounding boxes to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped bounding boxes. ##### apply_to_keypoints ```python def apply_to_keypoints(keypoints: np.ndarray, orig_width: int, **params) -> np.ndarray: ``` Flip keypoints horizontally and swap symmetric pairs. Parameters * `keypoints` (`np.ndarray`): Keypoints to flip. * `orig_width` (`int`): Original image width. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped keypoints. Raises * `ValueError`: If the total number of keypoints is not a multiple of `n_keypoints`. ##### apply_to_mask ```python def apply_to_mask(img: np.ndarray, **params) -> np.ndarray: ``` Flip a segmentation mask horizontally. Parameters * `img` (`np.ndarray`): Segmentation mask to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped segmentation mask. ##### get_params_dependent_on_data ```python def get_params_dependent_on_data(params: dict[str, Any], data: dict[str, Any]) -> dict[str, Any]: ``` Get parameters dependent on the targets. Parameters * `params` (`dict[str, Any]`): Existing augmentation parameters. * `data` (`dict[str, Any]`): Input data. Returns * `dict[str, Any]`: Parameters derived from the input targets. #### Attributes ##### keypoint_pairs Pairs of keypoint indices swapped after flipping. ##### n_keypoints Number of unique keypoints described by `keypoint_pairs`. ##### targets ### TransposeSymmetricKeypoints Transpose images and symmetric keypoints. #### Methods ##### init ```python def __init__(keypoint_pairs: list[tuple[int, int]], p: float = 0.5): ``` Transpose an image and symmetric keypoints. Equivalent to 90 degree rotation followed by horizontal flip. Parameters * `keypoint_pairs` (`list[tuple[int, int]]`): Pairs of keypoint indices to swap after the transpose. * `p` (`float`): Probability of applying the augmentation. ##### apply ```python def apply(img: np.ndarray, **params) -> np.ndarray: ``` Transpose an image. Parameters * `img` (`np.ndarray`): Image to transpose. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Transposed image. ##### apply_to_bboxes ```python def apply_to_bboxes(bboxes: np.ndarray, **params) -> np.ndarray: ``` Transpose bounding boxes. Parameters * `bboxes` (`np.ndarray`): Bounding boxes to transpose. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Transposed bounding boxes. ##### apply_to_keypoints ```python def apply_to_keypoints(keypoints: np.ndarray, **params) -> np.ndarray: ``` Transpose keypoints and swap symmetric pairs. Parameters * `keypoints` (`np.ndarray`): Keypoints to transpose. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Transposed keypoints. Raises * `ValueError`: If the total number of keypoints is not a multiple of `n_keypoints`. ##### apply_to_mask ```python def apply_to_mask(mask: np.ndarray, **params) -> np.ndarray: ``` Transpose a segmentation mask. Parameters * `mask` (`np.ndarray`): Segmentation mask to transpose. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Transposed segmentation mask. ##### get_params_dependent_on_data ```python def get_params_dependent_on_data(params: dict[str, Any], data: dict[str, Any]) -> dict[str, Any]: ``` Get parameters dependent on the targets. Parameters * `params` (`dict[str, Any]`): Existing augmentation parameters. * `data` (`dict[str, Any]`): Input data. Returns * `dict[str, Any]`: Parameters derived from the input targets. #### Attributes ##### keypoint_pairs Pairs of keypoint indices swapped after transposition. ##### n_keypoints Number of unique keypoints described by `keypoint_pairs`. ##### targets ### VerticalSymmetricKeypointsFlip Flip images and symmetric keypoints vertically. #### Methods ##### init ```python def __init__(keypoint_pairs: list[tuple[int, int]], p: float = 0.5): ``` Flip an image and symmetric keypoints vertically. Bounding boxes and segmentation masks are flipped as well. Parameters * `keypoint_pairs` (`list[tuple[int, int]]`): Pairs of keypoint indices to swap after the flip. * `p` (`float`): Probability of applying the augmentation. ##### apply ```python def apply(img: np.ndarray, **params) -> np.ndarray: ``` Flip an image vertically. Parameters * `img` (`np.ndarray`): Image to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped image. ##### apply_to_bboxes ```python def apply_to_bboxes(bboxes: np.ndarray, **params) -> np.ndarray: ``` Flip bounding boxes vertically. Parameters * `bboxes` (`np.ndarray`): Bounding boxes to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped bounding boxes. ##### apply_to_keypoints ```python def apply_to_keypoints(keypoints: np.ndarray, orig_height: int, **params) -> np.ndarray: ``` Flip keypoints vertically and swap symmetric pairs. Parameters * `keypoints` (`np.ndarray`): Keypoints to flip. * `orig_height` (`int`): Original image height. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped keypoints. Raises * `ValueError`: If the total number of keypoints is not a multiple of `n_keypoints`. ##### apply_to_mask ```python def apply_to_mask(img: np.ndarray, **params) -> np.ndarray: ``` Flip a segmentation mask vertically. Parameters * `img` (`np.ndarray`): Segmentation mask to flip. * `**params`: Additional transform parameters. Returns * `np.ndarray`: Flipped segmentation mask. ##### get_params_dependent_on_data ```python def get_params_dependent_on_data(params: dict[str, Any], data: dict[str, Any]) -> dict[str, Any]: ``` Get parameters dependent on the targets. Parameters * `params` (`dict[str, Any]`): Existing augmentation parameters. * `data` (`dict[str, Any]`): Input data. Returns * `dict[str, Any]`: Parameters derived from the input targets. #### Attributes ##### keypoint_pairs Pairs of keypoint indices swapped after flipping. ##### n_keypoints Number of unique keypoints described by `keypoint_pairs`. ##### targets