It’s looks like before your imports, there were only 15 encounters for this species in Flukebook. Historically, when we don’t have enough data to train a detector, we’ll let users submit encounters for a species so they can manually annotate them and start building a database we can work from. For the time being, you may need to manually annotate these imported encounters until I can get an update from the ML team on next steps.
Additionally, you have “St.Lucia” as your location ID in the spreadsheet for excelbulkimport32022Ks, but in should be written “St. Lucia” so it matches with the location ID name in Flukebook. This doesn’t affect detection, but when an algorithm is configured it will affect who is displayed as match candidates since Wildbook will treat it as a new, unique location.
Hi Anastasia,
I saw that when i want to put the species name in the “Select annotation iaClass” (Manual Annotation), I can choose between “Choose” and “pygmyspermwhale+fin_dorsal”. But the species I study is the dwarf sperm whale (Kogia sima) for all my bulk imports ending with “Ks.xlsx”
Is that normal ?
Should I re-import my bulk imports to reset the detection or not?
thank’s for your answer, I wish you to have a good day !
sincerely
Great question! That means that our existing pygmy sperm whale detector has also been configured for use on dwarf sperm whales. When species have similar characteristics, a single detector is cross-applied to multiple species. You don’t need to change the taxonomy in your spreadsheet.
You may want to delete and re-import the spreadsheet for excelbulkimport32022Ks because that was the one that needed the location ID updated for St. Lucia. Your imports have already been sent to detection (and some are ready for identification).
Let us know if you have any feedback about the algorithm’s performance!
Quick update: The IAclass names have been updated on the front end, so you should see the correct species name going forward. We also re-sent the imports to detection so they reflect the corrected class names.