https://github.com/ianhi/jupyter-microscopy

Tools for interacting/viewing microscopy data in Jupyter
https://github.com/ianhi/jupyter-microscopy

Last synced: 12 days ago
JSON representation

Tools for interacting/viewing microscopy data in Jupyter

• Host: GitHub
• URL: https://github.com/ianhi/jupyter-microscopy
• Owner: ianhi
• License: bsd-3-clause
• Created: 2020-07-13T00:18:30.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2020-09-13T03:29:56.000Z (over 4 years ago)
• Last Synced: 2024-11-05T08:03:05.494Z (about 2 months ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 4.87 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# jupyter-microscopy

Collection of tools for working with microscopy data in jupyter notebooks.

Outside of this repo some useful tools are:  

  

manual image segmentation: [ipysegment](https://github.com/ianhi/ipysegment)   

colony counting: [skimage counter](https://github.com/jrussell25/colony_counter)  

colony counting2: [web tool](https://github.com/ianhi/colony-counter)  

choose points of interest and output files of the points: [laser pointer](https://github.com/Hekstra-Lab/laser-pointer)

## Why?

I prefer python over ImageJ/FIJI for analyzing microscopy images, but I really miss the nice interactive image viewing provided imagej. The point of this repo is keep all the microscopy image tools I might make in a single place. I'll strive to replicate the most useful manual interaction tools provided by ImageJ in such a way to be useful to an advanced python user. Three key goals:

1. Limit distance to data

   - Should always be trivial to extract the data from manual interaction to a numpy array

2. Should *feel* good to use

   - I don't want my tools to make me sad, only my experiment not working gets to do that...

3. Doesn't look gross

   - similar to the above, so many people have thought hard about web design surely I can make use of some of that. 

   

   

## ROADMAP

Mostly created by looking through the options that FIJI provides and picking out the ones that seem useful and that aren't already sastified by something like scikit-image or just generic numpy functions

- [ ] Adjust image brightness and contrast

    - skimage auto isn't always great, get whatever FIJI auto does

- [ ] Image Stack viewer

    - Use [PIMS](https://github.com/soft-matter/pims) to store image data

- [ ] Region of interest selectors

    - ipysegment probably will be helpful for this. 

    - also see https://github.com/ideonate/jupyter-innotater for bounding boxes

- [ ] Basic image annotation (i.e. mspaint drawing)

    - surely this already exists?

Need to make sure that everything exposes the relevant variables in such a way that they can be `jslink`ed together in order to achieve goal 2. 

## Setting up a conda environment

I used the following bash commands to create envs. Add them to your `.bashrc`, do `source ~/.bashrc` and then run `jlab-env-full micro`

```bash

jlab-env-basic ()

{

    conda create -n $1 --override-channels --strict-channel-priority -c conda-forge -c anaconda jupyterlab nodejs python mamba -y

     conda activate $1

}

jlab-env-full ()

{  

  conda create -n $1 python -y

  conda activate $1

  conda install -c conda-forge mamba -y

  mamba install -c conda-forge jupyterlab nodejs scipy matplotlib numpy ipympl pandas -y

  pip install jupyterlab-git

  jupyter labextension install @jupyter-widgets/jupyterlab-manager --no-build

  jupyter lab build --name=$1

}

```