Mathieu Boudreau, PhD

@hermancollin Good call - we should replace the U-net image with that figure, as that one isn't going to be relevant soon correct?

nibabel pinned to 3.2.* on our end is in conflict with ivadomed, but the only reason we had it pinned to that version was to be on the same wavelength...

> Is this still a draft? Yes, I’m going convert it to point to the newest release of ivadomed now that it was made today!

> @mathieuboudreau not sure why I never reviewed this, sorry about that. I like this PR a lot and would like to revive it. We should merge this during/after the...

@apuer11 did you get answers to your questions / did you problem get resolved? Or did you close this issue accidentally?

Create an initial QA class for metrics, it saves histograms with some statistics next to them. Did an initial integration into Napari using the pre-existing Morphmetrics button (all metrics histograms...

One additional output I'd like to include are some suspected wrongly segmented axon IDs / locations, eg gratio = 0 or 1, axon_diam = 0, solidity = 1, etc. Maybe...

Adding on to this feature, I just made a proof-of-concept implementation of a feature that takes some pre-determined metric conditions to flag potential image objects that have issus (eg g-ratio...

How did we handle this in FSLeyes? I think if we changed images, it prompted us again, but I'm not 100% certain

>Also, in a fresh ADS install, the models are not downloaded at first. I think the reason we don't have any issue with users about this is that when the...