meghalithic / DeepBryo_micro

A web app for AI-assisted segmentation of cheilostome bryozoan colonies
Other
1 stars 0 forks source link

Holes and pebbles as orifices #13

Closed lilysandstrom closed 1 month ago

lilysandstrom commented 1 month ago

Holes, pebbles and other non-bryozoa are often detected as zooids with orifices, or just "disembodied" orifices. These generally have greater than 0.85 circularity (in M. speculum at least) and can be filtered out by that measure. See MHR.4858 for example: Screenshot from 2024-07-26 10-08-25

lilysandstrom commented 1 month ago

Also implementing a minimum & maximum percentage of the zooid that the orifice is allowed to take up (using knowledge of species), and/or a minimum & maximum absolute size for orifices will help remove falsly detected zooids with features. These rules can be used for other features as well, not only orifices; pebbles and holes are falsely detected as avicularia, ovicells and ascopores too. Eg. image MHR.5005 with orifice 17 double the correct size: Screenshot from 2024-07-26 11-01-24 Eg. image MHR.5053 with avicularium misidentified as very small orifice: image

lilysandstrom commented 1 month ago

Hole in image MHR.5240 identified as a small round orifice. Rotating the image does not help. image

lilysandstrom commented 1 month ago

Image MHR.5252 with misdetected avicularia nr 47 and 48. Screenshot from 2024-07-26 12-15-32

lilysandstrom commented 1 month ago

MHR.8292 has an example of a detected orifice that will be removed by maximum size limit: Screenshot from 2024-07-26 13-22-55

lilysandstrom commented 1 month ago

MHR.6487.jpg has holes detected as a large orifice on the right and small orifice on the left of the image: image

lilysandstrom commented 1 month ago

Image MHR.6553.jpg has an orifice that can be removed using maximum size/midth/height limit: image