dacapo.experiments.trainers.gp_augments

Submodules

• dacapo.experiments.trainers.gp_augments.augment_config
• dacapo.experiments.trainers.gp_augments.elastic_config
• dacapo.experiments.trainers.gp_augments.gamma_config
• dacapo.experiments.trainers.gp_augments.intensity_config
• dacapo.experiments.trainers.gp_augments.intensity_scale_shift_config
• dacapo.experiments.trainers.gp_augments.simple_config

Package Contents

Classes

AugmentConfig	Base class for gunpowder augment configurations. Each subclass of a Augment
ElasticAugmentConfig	A class that holds the configuration details for the elastic augmentations.
SimpleAugmentConfig	This class is an implementation of AugmentConfig that applies simple augmentations.
GammaAugmentConfig	This class manages the configuration of gamma augmentation for a given dataset.
IntensityAugmentConfig	Base class for gunpowder augment configurations. Each subclass of a Augment
IntensityScaleShiftAugmentConfig	Base class for gunpowder augment configurations. Each subclass of a Augment

class dacapo.experiments.trainers.gp_augments.AugmentConfig

Base class for gunpowder augment configurations. Each subclass of a Augment should have a corresponding config class derived from AugmentConfig.

abstract node(raw_key: gunpowder.ArrayKey, gt_key: gunpowder.ArrayKey, mask_key: gunpowder.ArrayKey) → gunpowder.BatchFilter

return a gunpowder node that performs this augmentation

class dacapo.experiments.trainers.gp_augments.ElasticAugmentConfig

A class that holds the configuration details for the elastic augmentations.

control_point_spacing

Distance(in voxels per dimension) between control points for the elastic deformation.

Type:

List[int]

control_point_displacement_sigma

Standard deviation of control point displacement distribution, in world coordinates.

Type:

List[float]

rotation_interval

An interval to randomly sample rotation angles from (0,2PI).

Type:

Tuple[float, float]

subsample

Downsample factor to perform the elastic augmentation on a grid. Default is 1.

Type:

int

uniform_3d_rotation

Should 3D rotations be performed uniformly. The ‘rotation_interval’ will be ignored. Default is False.

Type:

bool

control_point_spacing: List[int]

control_point_displacement_sigma: List[float]

rotation_interval: Tuple[float, float]

subsample: int

uniform_3d_rotation: bool

node(_raw_key=None, _gt_key=None, _mask_key=None)

Returns the object of ElasticAugment with the given configuration details.

Parameters:

• _raw_key – Unused variable, kept for future use.

• _gt_key – Unused variable, kept for future use.

• _mask_key – Unused variable, kept for future use.

Returns:

A ElasticAugment object configured with control_point_spacing,

control_point_displacement_sigma, rotation_interval, subsample and uniform_3d_rotation.

Return type:

ElasticAugment

class dacapo.experiments.trainers.gp_augments.SimpleAugmentConfig

This class is an implementation of AugmentConfig that applies simple augmentations.

Parameters:

• _raw_key – Key for raw data. Not used in this implementation. Defaults to None.

• _gt_key – Key for ground truth data. Not used in this implementation. Defaults to None.

• _mask_key – Key for mask data. Not used in this implementation. Defaults to None.

Returns:

A node that can be included in a pipeline to perform simple data augmentations.

Return type:

Gunpowder SimpleAugment Node

node(_raw_key=None, _gt_key=None, _mask_key=None)

Get a gp.SimpleAugment node.

Parameters:

• _raw_key ([type], optional) – Specific key for raw data, not used in this implementation. Defaults to None.

• _gt_key ([type], optional) – Specific key for ground truth data, not used in this implementation. Defaults to None.

• _mask_key ([type], optional) – Specific key for mask data, not used in this implementation. Defaults to None.

Returns:

Simple augmentation node which can be incorporated in the pipeline.

Return type:

gunpowder.SimpleAugment

class dacapo.experiments.trainers.gp_augments.GammaAugmentConfig

This class manages the configuration of gamma augmentation for a given dataset.

gamma_range

A tuple of float values represents the min and max range of gamma noise

to apply on the raw data.

node()

Constructs a node in the augmentation pipeline.

gamma_range: Tuple[float, float]

node(raw_key: gunpowder.ArrayKey, _gt_key=None, _mask_key=None)

Constructs a node in the augmentation pipeline.

Parameters:

• raw_key (gp.ArrayKey) – Key to an Array (volume) in the pipeline

• _gt_key (gp.ArrayKey, optional) – Ground Truth key, not used in this function. Defaults to None.

• _mask_key (gp.ArrayKey, optional) – Mask Key, not used in this function. Defaults to None.

Returns:

The augmentation method to be applied on the source data.

Return type:

GammaAugment instance

class dacapo.experiments.trainers.gp_augments.IntensityAugmentConfig

Base class for gunpowder augment configurations. Each subclass of a Augment should have a corresponding config class derived from AugmentConfig.

scale: Tuple[float, float]

shift: Tuple[float, float]

clip: bool

node(raw_key: gunpowder.ArrayKey, _gt_key=None, _mask_key=None)

return a gunpowder node that performs this augmentation

class dacapo.experiments.trainers.gp_augments.IntensityScaleShiftAugmentConfig

Base class for gunpowder augment configurations. Each subclass of a Augment should have a corresponding config class derived from AugmentConfig.

scale: float

shift: float

node(raw_key: gunpowder.ArrayKey, _gt_key=None, _mask_key=None)

return a gunpowder node that performs this augmentation