search

kevinjohncutler / omnipose

Omnipose: a high-precision solution for morphology-independent cell segmentation
https://omnipose.readthedocs.io
Other
85 stars 29 forks source link

Segmentation of C elegans #2

Closed LRpz closed 2 years ago

LRpz commented 2 years ago

Hi,

I was wondering whether the 'bact_omni' model could be used to segment C elegans grayscale images as shown in Figure 5. I only get oversegmented patches in some worms regardless of the diameter setting value.

Diameter=20 C_elegans_d=20

Diameter=250 C_elegans_d=250

Would you suggest changing model/training my own? Thanks a lot

kevinjohncutler commented 2 years ago

Hi @LRpz, sorry for the late reply. I have a model trained for c_elegans that I don't think is built in yet through cellpose. I'd be happy to send it to you (it is too big to host on github); send me an email at kcutler@uw.edu. Please share a raw example image too so that I can recommend the best settings as well.