frarch.datasets.transforms.pil_transforms module

class frarch.datasets.transforms.pil_transforms.PILBrightness(var: float)

Bases: object

Randomly dim or enhance brightness of an image.

Randomly dim or enhance brightness of an image given as an input. Given var value, enhance brightness from a 1-var factor to a 1+var factor.

Parameters

var (float) – float value to get random alpha value from [1-var, 1+var].

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_brightness = PILBrightness()
processed_image = random_brightness(image)

class frarch.datasets.transforms.pil_transforms.PILColorBalance(var: float)

Bases: object

Randomly dim or enhance color of an image.

Randomly dim or enhance color of an image given as an input. Given var value, enhance color from a 1-var factor to a 1+var factor.

Parameters

var (float) – float value to get random alpha value from [1-var, 1+var].

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_color_balance = PILColorBalance()
processed_image = random_color_balance(image)

class frarch.datasets.transforms.pil_transforms.PILContrast(var: float)

Bases: object

Randomly dim or enhance contrast of an image.

Randomly dim or enhance contrast of an image given as an input. Given var value, enhance contrast from a 1-var factor to a 1+var factor.

Parameters

var (float) – float value to get random alpha value from [1-var, 1+var].

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_contrast = PILContrast()
processed_image = random_contrast(image)

class frarch.datasets.transforms.pil_transforms.PILSharpness(var: float)

Bases: object

Randomly dim or enhance sharpness of an image.

Randomly dim or enhance sharpness of an image given as an input. Given var value, enhance sharpness from a 1-var factor to a 1+var factor.

Parameters

var (float) – float value to get random alpha value from [1-var, 1+var].

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_sharpness = PILSharpness()
processed_image = random_sharpness(image)

class frarch.datasets.transforms.pil_transforms.PowerPIL(rotate: bool = True, flip: bool = True, colorbalance: float = 0.4, contrast: float = 0.4, brightness: float = 0.4, sharpness: float = 0.4)

Bases: frarch.datasets.transforms.pil_transforms.RandomOrder

Composes several transforms together in random order.

Parameters

transforms (Iterable[Callable]) – An iterable containing callable objects which take PIL.Image as input and outputs the same object type.

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
powerpil = PowerPIL(True, True, 0.1, 0.1, 0.1, 0.1)
processed_image = powerpil(image)

class frarch.datasets.transforms.pil_transforms.RandomFlip

Bases: object

Randomly flips the given PIL.Image.

Probability of 0.25 horizontal flip, 0.25 vertical flip, 0.5 no flip.

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_flip = RandomFlip()
processed_image = random_flip(image)

class frarch.datasets.transforms.pil_transforms.RandomOrder(transforms: Optional[Iterable[Callable]])

Bases: object

Composes several transforms together in random order.

Parameters

transforms (Iterable[Callable]) – An iterable containing callable objects which take PIL.Image as input and outputs the same object type.

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_sharpness = RandomOrder(
    [
        RandomFlip(),
        RandomRotate(),
        PILColorBalance(0.1),
        PILContrast(0.1),
        PILBrightness(0.1),
        PILSharpness(0.1),
    ]
)
processed_image = random_sharpness(image)

class frarch.datasets.transforms.pil_transforms.RandomRotate

Bases: object

Randomly rotate the given PIL.Image.

Probability of 1/6 90°, 1/6 180°, 1/6 270°, 1/2 as is.

Example

Simple usage of the class:

image = PIL.Image.open("testimage.jpg")
random_rotate = RandomRotate()
processed_image = random_rotate(image)