Analysis tools & workflows of interest
Collection of potentially interesting tools or approaches for image analysis in the ENS
Some of the links point to the "Other Resources" pages of the previous sections.
Cell segmentation
Cellpose: State of the art cell segmentation software
In FIJI, using iterative thresholding has worked for me many times when segmenting neurons or enteric glial cells.
ilastik: Using a pixel classifier for segmentation, or using it in combination with object classifier for classifying cells
3D segmentation
Cellpose
Cellpose can be used for 3D segmentation of neurons or other cell types. To create a model that can work on 3D data:
Prepare your data:
Convert your 3D dataset into 2D views for training. You'll need to generate three types of 2D images: top (XY), side (XZ), and front (YZ) views.
To do this conversion, you could use the "Reslice" feature in FIJI (Image -> Stack -> Reslice). This will create different views of your dataset, from which you can select a 2D slice.
Annotate the Views:
Once you have your 2D images (XY, XZ, YZ views), you'll need to label or annotate them. You could do this in cellpose.
Train Your Model:
With your annotated 2D views ready, you can now proceed to train your model. It is recommended to try the default models and assess which works best on your data. Use this model for finetuning on yoru dataset
Vollseg
Vollseg allows segmentation of objects with irregular shapes. It does require configuration in Python.
Last updated