Transforms

Data transformations and augmentations for 2D and 3D BioImages

Sampling

---

source

Resample

 Resample (sampling, **kwargs)

*A subclass of Spacing that handles image resampling based on specified sampling factors or voxel dimensions.

The Resample class inherits from Spacing and provides a flexible way to adjust the spacing (voxel size) of images by specifying either a sampling factor or explicitly providing new voxel dimensions.

Args: sampling (int, optional): Sampling factor for isotropic resampling. Default is 1, indicating no change in resolution. **kwargs: Additional keyword arguments that can include ‘pixdim’ to specify custom voxel dimensions.

Attributes: pixdim (list or tuple): The voxel dimensions of the image after resampling. If not provided during initialization, this will be determined based on the sampling factor and original image properties.*

from bioMONAI.core import cells3d, img2Tensor
from bioMONAI.visualize import visualize_slices

img = BioImageStack(img2Tensor(cells3d()[:,0]))
visualize_slices(img, showlines=False)

img2 = Resample(4)(img)
visualize_slices(img2, showlines=False)

Noise

---

source

RandCameraNoise

 RandCameraNoise (p:float=1.0, damp=0.01, qe=0.7, gain=2, offset=100,
                  exp_time=0.1, dark_current=0.6, readout=1.5,
                  bitdepth=16, seed=42, simulation=False, camera='cmos',
                  gain_variance=0.1, offset_variance=5)

*Simulates camera noise by adding Poisson shot noise, dark current noise, and optionally CMOS fixed pattern noise.

Returns: numpy.ndarray: The noisy image as a NumPy array with dimensions of input_image.*

	Type	Default	Details
p	float	1.0	Probability of applying Transform
damp	float	0.01	Dampening factor to prevent saturation when adding noise
qe	float	0.7	Quantum efficiency of the camera (0 to 1).
gain	int	2	Camera gain factor. If an array, it should be broadcastable with input_image shape.
offset	int	100	Camera offset in ADU. If an array, it should be broadcastable with input_image shape.
exp_time	float	0.1	Exposure time in seconds.
dark_current	float	0.6	Dark current per pixel in electrons/second.
readout	float	1.5	Readout noise standard deviation in electrons.
bitdepth	int	16	Bit depth of the camera output.
seed	int	42	Seed for random number generator for reproducibility.
simulation	bool	False	If True, assumes input_image is already in units of photons and does not convert from electrons.
camera	str	cmos	Specifies the type of camera (‘cmos’ or any other). Used to add CMOS fixed pattern noise if ‘cmos’ is specified.
gain_variance	float	0.1	Variance for the gain noise in CMOS cameras. Only applicable if camera type is ‘cmos’.
offset_variance	int	5	Variance for the offset noise in CMOS cameras. Only applicable if camera type is ‘cmos’.

from bioMONAI.visualize import plot_image

img3 = img[30]

plot_image(img3)
plot_image(RandCameraNoise(camera = 'cmos').encodes(img3))
plot_image(RandCameraNoise(camera = 'ccd', readout=2).encodes(img3))

Normalization

---

source

ScaleIntensity

 ScaleIntensity (x, min=0.0, max=1.0, axis=None, eps=1e-20, dtype=<class
                 'numpy.float32'>)

Percentile-based image normalization.

---

source

ScaleIntensityPercentiles

 ScaleIntensityPercentiles (x, pmin=3, pmax=99.8, axis=None, eps=1e-20,
                            dtype=<class 'numpy.float32'>)

Percentile-based image normalization.

---

source

ScaleIntensityVariance

 ScaleIntensityVariance (ndim=2)

*Scales the intensity variance of an ND image to a target value.

Args: target_variance (float): The desired variance for the scaled intensities. ndim (int): Number of spatial dimensions in the image. Default is 2 for 2D images.*

# Example usage with a random tensor of shape (1, 3, 256, 256)
rand_tensor = BioImageBase(torch.rand(1, 3, 256, 256))

transform = ScaleIntensityVariance(ndim=4)

# Apply the transform to the tensor
scaled_tensor = transform(rand_tensor)

print('Original Tensor Variance:', rand_tensor.var().item())
print('Scaled Tensor Variance:', scaled_tensor.var().item())

Original Tensor Variance: 0.08341076225042343
Scaled Tensor Variance: 1.0

Data Augmentation

---

source

RandCrop2D

 RandCrop2D (size:int|tuple, lazy=False, **kwargs)

Randomly crop an image to size

	Type	Default	Details
size	int | tuple		Size to crop to, duplicated if one value is specified
lazy	bool	False	a flag to indicate whether this transform should execute lazily or not. Defaults to False
kwargs

---

source

RandCropND

 RandCropND (size:int|tuple, lazy=False, **kwargs)

*Randomly crops an ND image to a specified size.

This transform randomly crops an ND image to a specified size during training and performs a center crop during validation. It supports both 2D and 3D images and videos, assuming the first dimension is the batch dimension.

Args: size (int or tuple): The size to crop the image to. this can have any number of dimensions. If a single value is provided, it will be duplicated for each spatial dimension, up to a maximum of 3 dimensions. **kwargs: Additional keyword arguments to be passed to the parent class.*

	Type	Default	Details
size	int | tuple		Size to crop to, duplicated if one value is specified
lazy	bool	False	a flag to indicate whether this transform should execute lazily or not. Defaults to False
kwargs

# Define a random tensor
orig_size = (65, 65)
rand_tensor = BioImageBase(torch.rand(8, *orig_size))

for i in range(100):
    test_eq((8,64,64),RandCropND((64,64))(rand_tensor).shape)

---

source

RandFlip

 RandFlip (prob=0.1, spatial_axis=None, ndim=2, lazy=False, **kwargs)

Randomly flips an ND image over a specified axis.

	Type	Default	Details
prob	float	0.1	Probability of flipping
spatial_axis	NoneType	None	Spatial axes along which to flip over. Default is None. The default axis=None will flip over all of the axes of the input array.
ndim	int	2	If axis is negative it counts from the last to the first axis. If axis is a tuple of ints, flipping is performed on all of the axes specified in the tuple.
lazy	bool	False	Flag to indicate whether this transform should execute lazily or not. Defaults to False
kwargs

# Define a random tensor
orig_size = (1,4,4)
rand_tensor = BioImageBase(torch.rand(*orig_size))

print('orig tensor: ', rand_tensor, '\n')

for i in range(3):
    print(RandFlip(prob=.75, spatial_axis=None)(rand_tensor))

orig tensor:  metatensor([[[0.8943, 0.3943, 0.2556, 0.8066],
         [0.5016, 0.6177, 0.8749, 0.1229],
         [0.9610, 0.1088, 0.5785, 0.6857],
         [0.5776, 0.7028, 0.5421, 0.3654]]]) 

metatensor([[[0.8066, 0.2556, 0.3943, 0.8943],
         [0.1229, 0.8749, 0.6177, 0.5016],
         [0.6857, 0.5785, 0.1088, 0.9610],
         [0.3654, 0.5421, 0.7028, 0.5776]]])
metatensor([[[0.8943, 0.3943, 0.2556, 0.8066],
         [0.5016, 0.6177, 0.8749, 0.1229],
         [0.9610, 0.1088, 0.5785, 0.6857],
         [0.5776, 0.7028, 0.5421, 0.3654]]])
metatensor([[[0.8066, 0.2556, 0.3943, 0.8943],
         [0.1229, 0.8749, 0.6177, 0.5016],
         [0.6857, 0.5785, 0.1088, 0.9610],
         [0.3654, 0.5421, 0.7028, 0.5776]]])

---

source

RandRot90

 RandRot90 (prob=0.1, max_k=3, spatial_axes=(0, 1), ndim=2, lazy=False,
            **kwargs)

Randomly rotate an ND image by 90 degrees in the plane specified by axes.

	Type	Default	Details
prob	float	0.1	Probability of rotating
max_k	int	3	Max number of times to rotate by 90 degrees
spatial_axes	tuple	(0, 1)	Spatial axes along which to rotate. Default: (0, 1), this is the first two axis in spatial dimensions.
ndim	int	2
lazy	bool	False	Flag to indicate whether this transform should execute lazily or not. Defaults to False
kwargs

# Define a random tensor
orig_size = (1,4,4)
rand_tensor = BioImageBase(torch.rand(*orig_size))

print('orig tensor: ', rand_tensor, '\n')

for i in range(3):
    print(RandRot90(prob=.75)(rand_tensor))

orig tensor:  metatensor([[[0.3961, 0.2351, 0.4363, 0.3366],
         [0.9416, 0.8529, 0.8646, 0.0091],
         [0.5012, 0.9804, 0.8109, 0.1094],
         [0.9639, 0.5107, 0.4666, 0.8632]]]) 

metatensor([[[0.3961, 0.2351, 0.4363, 0.3366],
         [0.9416, 0.8529, 0.8646, 0.0091],
         [0.5012, 0.9804, 0.8109, 0.1094],
         [0.9639, 0.5107, 0.4666, 0.8632]]])
metatensor([[[0.9639, 0.5012, 0.9416, 0.3961],
         [0.5107, 0.9804, 0.8529, 0.2351],
         [0.4666, 0.8109, 0.8646, 0.4363],
         [0.8632, 0.1094, 0.0091, 0.3366]]])
metatensor([[[0.8632, 0.4666, 0.5107, 0.9639],
         [0.1094, 0.8109, 0.9804, 0.5012],
         [0.0091, 0.8646, 0.8529, 0.9416],
         [0.3366, 0.4363, 0.2351, 0.3961]]])