Affine

class deeptrack.augmentations.Affine(scale: PropertyLike[float] = 1, translate: PropertyLike[float | None] = None, translate_px: PropertyLike[float] = 0.0, rotate: PropertyLike[float] = 0.0, shear: PropertyLike[float] = 0.0, order: PropertyLike[int] = 1, cval: PropertyLike[float] = 0.0, mode: PropertyLike[str] = 'reflect', **kwargs)

Bases: Augmentation

Augmenter to apply affine transformations to images.

Affine transformations include:

• Translation

• Scaling

• Rotation

• Shearing

Some transformations involve interpolations between several pixels of the input image to generate output pixel values. The parameter order deals with the method of interpolation used for this.

Parameters

scale: float or tuple of floats or list of floats or dict

Scaling factor to use, where 1.0 denotes “no change” and 0.5 is zoomed out to 50 percent of the original size. If two values are provided (using tuple, list, or dict), the two first dimensions of the input are scaled individually.

translate: float or tuple of floats or list of floats or dict

Translation in pixels.

translate_px: float or tuple of floats or list of floats or dict

DEPRECATED, use translate.

rotate: float

Rotation in radians, i.e. Rotation happens around the center of the image.

shear: float

Shear in radians. Values in the range (-pi/4, pi/4) are common.

order: int

Interpolation order to use. Same meaning as in skimage:

• 0: Nearest-neighbor

• 1: Bi-linear (default)

• 2: Bi-quadratic (not recommended by skimage)

• 3: Bi-cubic

• 4: Bi-quartic

• 5: Bi-quintic

cval: float

The constant intensity value used to fill in new pixels. This value is only used if mode is set to constant.

mode: str

Parameter that defines newly created pixels. May take the same values as in scipy.ndimage.affine_transform(), i.e. constant, nearest, reflect or wrap.

Methods

_process_properties(properties: dict) -> dict

Processes the properties of the image.

get(image: Image | np.ndarray, scale: PropertyLike[float], translate: PropertyLike[float], rotate: PropertyLike[float], shear: PropertyLike[float], **kwargs) -> Image

Abstract method which performs the Affine augmentation.

Methods Summary

get(image, scale, translate, rotate, shear, ...)

Abstract method which performs the Affine augmentation.

Methods Documentation

get(image: Image | np.ndarray, scale: PropertyLike[float], translate: PropertyLike[float], rotate: PropertyLike[float], shear: PropertyLike[float], **kwargs) → Image

Abstract method which performs the Affine augmentation.

Affine transformations include: - Translation - Scaling - Rotation - Shearing