Clean Morphology Data?

[HaleySchuhl]

Start a discussion about anything you would like Morphological measurements are inherently verbose and volatile to imperfections in binary masks or other weirdness that happens with iterative thinning. We can can potentially clean up observations as a type of quality control step to get rid of obvious oddity segments that could skew mean segment length/angle/etc. Segment length is straightforward to clean. Are there other measurements that could be filtered reasonably?