cyto_dl.image.transforms.track_transforms module#

class cyto_dl.image.transforms.track_transforms.CropResize(keys: str | List | ListConfig, max_shift=8, allow_missing_keys: bool = False)[source]#

Bases: RandomizableTransform

Parameters:

  • keys (list) – List of keys to apply transform to

  • max_shift (int) – Maximum number of pixels to shift image by before resizing

  • allow_missing_keys (bool) – Whether to allow missing keys

class cyto_dl.image.transforms.track_transforms.GenerateTrackLabels(img_key: str = 'img', formation_key: str = 'formation', breakdown_key: str = 'breakdown', track_start_key: str = 'track_start', label_key: str = 'label')[source]#

Bases: Transform

Transform to generate track labels from breakdown and formation labels.

Parameters:

  • img_key (str) – Key with image

  • formation_key (str) – Key with formation

  • breakdown_key (str) – Key with breakdown

  • track_start_key (str) – Key with track start

  • label_key (str) – Key to save label into

class cyto_dl.image.transforms.track_transforms.PerChannel(keys: str | List | ListConfig, transform: Transform, allow_missing_keys: bool = False)[source]#

Bases: Transform

Transform to apply same transform to each channel of image.

Parameters:

  • keys (list) – List of keys to apply transform to

  • transform (Transform) – Transform to apply to each channel

  • allow_missing_keys (bool) – Whether to allow missing keys

class cyto_dl.image.transforms.track_transforms.SplitTrackd(img_key: str = 'img', label_key: str = 'label')[source]#

Bases: Transform

Parameters:

  • img_key (str) – Key with image

  • label_key (str) – Key with label