opt2mesh: Extract meshes from Optical Projective Tomography scans

Installation step

• Clone the repository.
• Put the models in the models folder
  • Some models which have been pretrained (from OPT scans of embryos for the UNetPipeline and the UNet3DPipeline are available here: https://oc.ebi.ac.uk/s/mVuDrBtROqTadhc)
• Create a conda virtual environment and install the package:

$ conda env create -f environment.yml
$ conda activate opt2mesh
$ python setup.py install

CLI Usage

You can access the method Simply run:

$ opt2mesh opt_scan.tif /path/to/result/folder

This will create a subfolder containing:

• the mesh extracted (opt_scan.stl)
• logs of the job ran
• information about the mesh correctness and quality

You can store the result of the segmentation by using:

$ opt2mesh --save_occupancy_map opt_scan.tif

All the options are available using the help flag:

$ opt2mesh -h

⚠ If you get:

ModuleNotFoundError: No module named 'opt2mesh'

when running opt2mesh directly, just use:

$ python opt2mesh

Segmentation methods

Several methods for segmentation are available to identify the object in the image. See the --method flag:

• acwe uses morphological variant of Active Contours Without Edge.
• gac uses morphological variant of Geodesic Active Contours.
• 2d_unet uses a 2D U-Net trained on this implementation.
• 3d_unet uses a 3D U-Net trained on this implementation. You definitely require a GPU for this method.
• direct performs a direct meshing on a already OPT scan. This is the default one method.

Each of those methods has its own set of parameters which can be set on the command line. See the help for more detail:

$ opt2mesh -h

Introducing a custom segmentation method

You can introduce your own segmentation method in the pipeline.

To do this, you have to extend the base class OPT2MeshPipeline and to define the _extract_occupancy_map method.