Skip to content

Albumentations Experimental Transforms (augmentations.transforms)

class albumentations_experimental.augmentations.transforms.FlipSymmetricKeypoints (symmetric_keypoints_horizontal=(), symmetric_keypoints_vertical=(), symmetric_keypoints_both=(), *args, **kwargs) [view source on GitHub]

Flip the input horizontally around the y-axis.

Parameters:

Name Type Description
p float

probability of applying the transform. Default: 0.5.

symmetric_keypoints_horizontal tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if horizontal flip swaps their semantics, e.g. left arm - right arm. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

symmetric_keypoints_vertical tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if vertical flip swaps their semantics, e.g. top corner - bottom corner. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

symmetric_keypoints_both tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if vertical and horizontal flip swaps their semantics, e.g. top left corner - bottom right corner. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

Targets: image, mask, bboxes, keypoints

Image types: uint8, float32

class albumentations_experimental.augmentations.transforms.HorizontalFlipSymmetricKeypoints (symmetric_keypoints, *args, **kwargs) [view source on GitHub]

Flip the input horizontally around the y-axis.

Parameters:

Name Type Description
p float

probability of applying the transform. Default: 0.5.

symmetric_keypoints tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if horizontal flip swaps their semantics, e.g. left arm - right arm. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

Targets: image, mask, bboxes, keypoints

Image types: uint8, float32

class albumentations_experimental.augmentations.transforms.TransposeSymmetricKeypoints (symmetric_keypoints=(), *args, **kwargs) [view source on GitHub]

Flip the input horizontally around the y-axis.

Parameters:

Name Type Description
p float

probability of applying the transform. Default: 0.5.

symmetric_keypoints tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if vertical and horizontal flip swaps their semantics, e.g. top left corner - bottom right corner. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

Targets: image, mask, bboxes, keypoints

Image types: uint8, float32

class albumentations_experimental.augmentations.transforms.VerticalFlipSymmetricKeypoints (symmetric_keypoints, *args, **kwargs) [view source on GitHub]

Flip the input vertically around the x-axis.

Parameters:

Name Type Description
p float

probability of applying the transform. Default: 0.5.

symmetric_keypoints tuple, list, dict

tuple of pairs containing indices of symmetric keypoints. Keypoints are considered as symmetric if vertical flip swaps their semantics, e.g. top corner - bottom corner. If keypoint does not have pair then set as it is own index. Kyepoints length must be divisible to symmetric count. For example 5 symmetric pairs and 15 keypoints.

Targets: image, mask, bboxes, keypoints

Image types: uint8, float32