First of all, thnaks for this tool, it's really amazing.
I started using it for FIB-SEM dataset segmentation and it works great, at least for some structure such as mitochondria.
Would it be possible to remove some annotation, since that right now, if you do one extra point of the wrong type (e.g., positive instead of negative), it seems there is no way to remove it and restart the prediction ?
Best,
Benjamin