maurosilber / smo

Robust and unbiased estimation of the background distribution for fluorescence microscopy.

SMO

SMO is a Python package that implements the Silver Mountain Operator (SMO), which allows to recover an unbiased estimation of the background intensity distribution in a robust way.

We provide an easy to use Python package and plugins for some of the major image processing softwares: napari, CellProfiler, and ImageJ / FIJI. See Plugins section below.

Citation

To learn more about the theory behind SMO, you can read:

• the peer-reviewed article in the Journal of the Optical Society of America,
• the pre-print in BioRxiv.

If you use this software, please cite the peer-reviewed article.

Usage

To obtain a background-corrected image, it is as straightforward as:

import skimage.data
from smo import SMO

image = skimage.data.human_mitosis()
smo = SMO(sigma=0, size=7, shape=(1024, 1024))
background_corrected_image = smo.bg_corrected(image)

where we used a sample image from scikit-image. By default, the background correction subtracts the median value of the background distribution. Note that the background regions will end up with negative values, but with a median value of 0.

A notebook explaining in more detail the meaning of the parameters and other possible uses for SMO is available here: smo/examples/usage.ipynb .

Installation

It can be installed with pip from PyPI:

pip install smo

or with conda from the conda-forge channel:

conda install -c conda-forge smo

Plugins

Napari

A napari plugin is available.

To install:

• Option 1: in napari, go to Plugins > Install/Uninstall Plugins... in the top menu, search for smo and click on the install button.

• Option 2: just pip install this package in the napari environment.

It will appear in the Plugins menu.

CellProfiler

A CellProfiler plugin in available in the smo/plugins/cellprofiler folder.

To install, save this file into your CellProfiler plugins folder. You can find (or change) the location of your plugins directory in File > Preferences > CellProfiler plugins directory.

ImageJ / FIJI

An ImageJ / FIJI plugin is available in the smo/plugins/imagej folder.

To install, download this file and:

• Option 1: in the ImageJ main window, click on Plugins > Install... (Ctrl+Shift+M), which opens a file chooser dialog. Browse and select the downloaded file. It will prompt to restart ImageJ for changes to take effect.

• Option 2: copy into your ImageJ plugins folder (File > Show Folder > Plugins).

To use the plugin, type smo on the bottom right search box:

select smo in the Quick Search window and click on the Run button.

Note: the ImageJ plugin does not check that saturated pixels are properly excluded.

Development

Code style is enforced via pre-commit hooks. To set up a development environment, clone the repository, optionally create a virtual environment, install the [dev] extras and the pre-commit hooks:

git clone https://github.com/maurosilber/SMO
cd SMO
conda create -n smo python pip numpy scipy
pip install -e .[dev]
pre-commit install

smo's Issues

Question about details / theory

Hi Mauro, awesome package.

Do you have any material/reference about the Silver Mountain Operator?

Thanks