OmniSegger is the Supersegger MATLAB-based suite modified to work with improved Omnipose segmentation. Omnipose should be installed before running OmniSegger.
More information about Omnipose can be found at the Omnipose Github and documentation page.
- Software Requirements
- Software Documentation
- Installation Instructions
- Setting the Path
- Running OmniSegger (GUI)
- Running OmniSegger from MATLAB (no GUI) - Recommended!
- Saving clist as Excel file
- ND2 to TIFF
- Running Omnipose directly from MATLAB
- Segmentation of alternative imaging modalities
- Troubleshooting & Known Issues
OmniSegger uses the same MATLAB toolboxes as the original Supersegger:
- Curve Fitting Toolbox
- Deep Learning Toolbox (fka Neural Network Toolbox)
- Global Optimization Toolbox
- Image Processing Toolbox
- Optimization Toolbox
- Parallel Computing Toolbox (not necessary)
- Statistics and Machine Learning Toolbox
The GitHub for the original Supersegger is here. For more detailed documentation, the website for Supersegger can be found here, the wiki, and detailed documentation on functions found here. OmniSegger uses the same MATLAB functions as the original Supersegger.
Quick-start guide for new users \ Original SuperSegger guide to segmentation \ Viewing the results \ The clist
Omnipose options have been preselected to work directly with OmniSegger, but if needed, suggestions for segmentation options are discussed here. Recommended options can also be found on the documentation page.
git clone https://github.com/tlo-bot/supersegger-omnipose.git
- Add OmniSegger to MATLAB path, with its subfolders (see "Setting the Path").
- Install Omnipose:
In order for Matlab to be able to find OmniSegger, the OmniSegger folder needs to be in your path*. In the Home tab, in the Environment section, click Set Path. The Set Path dialog box appears. Click 'add folder with subfolders' and add the OmniSegger folder.
*note that if the original Supersegger is already installed and on the MATLAB path, you should replace the paths of the original Supersegger folders & subfolders with the paths to the new OmniSegger folders & subfolders.
- Put images (.tif) into a folder.
- Convert image file names to OmniSegger convention (in MATLAB with
superSeggerGui, orconvertImageNames; or manually with command line). - Run
superSeggerGuiwith the desired parameters. OmniSegger will begin aligning the images.- After aligning, OmniSegger will pause for segmentation through Omnipose. The Omnipose command should be displayed on the MATLAB Command Window and also automatically copied to your clipboard.
- Open terminal/Anaconda Prompt and activate the Omnipose environment (ie
conda activate omnipose)- Note that you should see the conda environment change from "(base)" to "(omnipose)"
- Paste in the Omnipose command that was generated by MATLAB into the terminal. Wait for Omnipose to segment images and generate masks.
- Can also change the segmentation command options in this step, if needed. For example, append
--use_gputo utilize the GPU.
- Can also change the segmentation command options in this step, if needed. For example, append
- Once Omnipose has completed, continue running OmniSegger by pressing the return/Enter key in the MATLAB Command Window.
- Configure the analysis parameters in MATLAB:
edit processExp. In particular, the 'Calculation Options' section should be edited depending on the desired fluorescence analysis. If not concerned with fluorescence, processExp probably doesn't need to be changed. - Run
processExp('dirname')to begin running the segmentation, where 'dirname' is the path to your image files. OmniSegger will begin aligning the images.- After aligning, OmniSegger will pause for segmentation through Omnipose. The Omnipose command should be displayed on the MATLAB Command Window and also automatically copied to your clipboard.
- Open terminal/Anaconda Prompt and activate the Omnipose environment (ie
conda activate omnipose)- Note that you should see the conda environment change from "(base)" to "(omnipose)"
- Paste in the Omnipose command that was generated by MATLAB into the terminal. Wait for Omnipose to segment images and generate masks.
- Can also change the segmentation command options in this step, if needed. For example, append
--use_gputo utilize the GPU.
- Can also change the segmentation command options in this step, if needed. For example, append
- Once Omnipose has completed, continue running OmniSegger by pressing the return/Enter key in the MATLAB Command Window.
Only supported when using processExp. Specify savexls=1 input for processExp (ie processExp('dirname',1)). Disabled by default.
If the clist has already been generated, running clist2xls('path/clist.mat') will save the clist as an xls in the xy directory.
Convert ND2 to individual TIFF images in OmniSegger naming format, and save out the metadata.
Dependencies can be installed to the Omnipose conda environment via: pip install aicsimageio[nd2] bioformats-jar
Usage: python /path/to/OmniSegger/SuperSegger-master/Internal/nd2totiff.py /path/to/your/file.nd2
Only supported when using processExp. Specify autoomni=1 input for processExp (ie processExp('dirname',[],1)). Disabled by default.
See documentation for setup linked below.
Coming soon!