Segmentation Tutorial Part 6 - veeninglab/BactMAP GitHub Wiki

To get answers to what is so different about the segmentations, I’ll go back to viewing the segmentations using plotRaw(). I think it really helps to see these outputs in the same way, as opposed to in the different segmentation viewers. Viewing them in one go makes it much easier to compare by eye.

You can plot all plots the way as we have done before for Clement’s segmentation, but I’m a lazy person and I know that I now have one file with the information of all segmentations combined (allFrames$finalframe). One small ggplot-command called facet_wrap can split up a plot by one or more variables, making it very straightforward to plot different conditions alongside each other.

plotRaw(tiffdata =pneumoTIFF, meshdata= allFrames$finalframe, xrange=c(800,1000),yrange=c(1400,1600)) +
        facet_wrap(~condition)

I think we cannot escape the fact that there is a big difference in the segmentations of the different programs (and people!). Before speculating on why, I’ll do some simple statistical tests.