search

kevinjohncutler / omnipose

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

Masks vs cells #26

Closed keara closed 1 year ago

keara commented 1 year ago

Omnipose seems to do really well segmented our bacteria with no additional training. I am a bit confused by the mask number as I assumed mask number would correspond to cell number. For example, I get 5 masks and 1712 cells. If I try to remove a cell in the GUI it removes all masks that are the same color. If I load the _seg.npy file and look at masks there are indeed only 5. I was expecting each cell to have a unique label. Is this the expected output? Thanks!

kevinjohncutler commented 1 year ago

@keara that is expected with the "ncolor" toggle on :) you can toggle it off to edit and save masks with the usual convention of 1,2,...,# of cells. I personally prefer to view with ncolor on, but you are right that some editing features are affected as you would expect with many cells receiving the same label. Glad it is working well for you!

kevinjohncutler commented 1 year ago

More details on ncolor are on the Docs too (just fixed the missing images): https://omnipose.readthedocs.io/ncolor.html

keara commented 1 year ago

Thanks for the clarification and speedy reply. That all makes sense - I hadn't fully read the ncolor part in the Docs, so my mistake for missing that.