from typing import Dict, Optional
import torch
from kornia.augmentation._2d.intensity.base import IntensityAugmentationBase2D
[docs]class RandomGaussianNoise(IntensityAugmentationBase2D):
r"""Add gaussian noise to a batch of multi-dimensional images.
.. image:: _static/img/RandomGaussianNoise.png
Args:
mean: The mean of the gaussian distribution.
std: The standard deviation of the gaussian distribution.
return_transform: if ``True`` return the matrix describing the transformation applied to each
input tensor. If ``False`` and the input is a tuple the applied transformation won't be concatenated.
same_on_batch: apply the same transformation across the batch.
p: probability of applying the transformation.
keepdim: whether to keep the output shape the same as input (True) or broadcast it
to the batch form (False).
Examples:
>>> rng = torch.manual_seed(0)
>>> img = torch.ones(1, 1, 2, 2)
>>> RandomGaussianNoise(mean=0., std=1., p=1.)(img)
tensor([[[[ 2.5410, 0.7066],
[-1.1788, 1.5684]]]])
To apply the exact augmenation again, you may take the advantage of the previous parameter state:
>>> input = torch.randn(1, 3, 32, 32)
>>> aug = RandomGaussianNoise(mean=0., std=1., p=1.)
>>> (aug(input) == aug(input, params=aug._params)).all()
tensor(True)
"""
def __init__(
self,
mean: float = 0.0,
std: float = 1.0,
return_transform: bool = False,
same_on_batch: bool = False,
p: float = 0.5,
keepdim: bool = False,
) -> None:
super().__init__(
p=p, return_transform=return_transform, same_on_batch=same_on_batch, p_batch=1.0, keepdim=keepdim
)
self.flags = dict(mean=mean, std=std)
def generate_parameters(self, shape: torch.Size) -> Dict[str, torch.Tensor]:
noise = torch.randn(shape)
return dict(noise=noise)
def apply_transform(
self, input: torch.Tensor, params: Dict[str, torch.Tensor], transform: Optional[torch.Tensor] = None
) -> torch.Tensor:
return input + params["noise"].to(input.device) * self.flags["std"] + self.flags["mean"]