oneVoxel

Small library for adding well described noise to images, with a command-line utility specifically for application on Nifti formatted images.

Installation

Simple! Just open your favourite terminal and type:

$ pip install onevox

Alongside installing the oneVoxel package, this will also ensure the dependencies are installed: numpy, scipy, nibabel, and nilearn.

For building this in Docker, you can run the following command:

$ docker build -t gkiar/onevox:local --network host .

Usage

From within Python this library can used to apply noise to arbitrary images:

[1]: import numpy as np
[2]: np.random.seed(1234)
[3]: data = np.random.random((10,10,10))  # Create data matrix
[4]: mask = data[:,:,0] > 0.4  # Define mask as values higher than 0.4 in the first 2D slice

[5]: from onevoxel import noise  # Load noise utils

[6]: # Generate noise locations from image and mask
[7]: loc = noise.generate_noise_params(data, mask, erode=0, mode='independent')
[8]: loc
[(3, 6, 0),
 (4, 4, 1),
 (9, 0, 2),
 (7, 9, 3),
 (1, 2, 4),
 (3, 1, 5),
 (0, 7, 6),
 (9, 0, 7),
 (3, 4, 8),
 (1, 9, 9)]

[9]: # Apply noise to the image and verify it's in the right spot
[10]: noisy_data, noisy_hash = noise.apply_noise_params(data, loc, scale=True, intensity=0.01)
[11]: sorted(list(zip(*np.where(noisy_data != data))), key=lambda elem: elem[2]) == loc
True

Contributing

Excited by the project and want to get involved?! Please check out our contributing guide, and look through the issues to start seeing where you can lend a hand. We look forward to approving your amazing contributions!