API reference

deepposekit.augment

Helpers that bridge the imgaug augmentation library with DeepPoseKit's training generators.

FlipAxis

Mirrors images along the X or Y axis and swaps keypoint indices accordingly, so symmetric body parts (e.g. left_wing↔ right_wing) stay consistent.

from deepposekit.augment import FlipAxis

flip = FlipAxis(generator=data_generator, axis=1)

Composing with imgaug

from imgaug import augmenters as iaa

augmenter = iaa.Sequential([
    flip,
    iaa.Affine(scale=(0.8, 1.2), rotate=(-180, 180)),
])

Once composed, pass the augmenter to a TrainingGenerator — see the augmentation guide for a recommended starting pipeline.