Sharkbook bug - comparing left flank to right flanks

In which Wildbook did the issue occur? - Sharkbook.ai

What operating system were you using? (eg. MacOS 10.15.3) - windows

What web browser were you using? (eg. Chrome 79) - firefox

What is your role on the site? (admin, researcher, etc) - shark researcher

What happened? on a few recent spot mappings it is showing me results of right flanks for a left flank shark. When I have checked the images it is providing as options the metadata is correctly saying that one is a left flank and the other is a right flank - yet they are showing up as potential matches.

What did you expect to happen? - left flank sharks should only match with left flank sharks. right flanks should only match with right flanks.

What are some steps we could take to reproduce the issue? review the potential matches shown here: Sharkbook: Wildbook for Sharks

Thanks for letting us know! I’ll need more time to research this. I’ll follow up when I have an update.

1 Like

Here is another good example.

The shark encounter is here: https://www.sharkbook.ai/encounters/encounter.jsp?number=be0dea0d-76c2-4c81-bb70-a1e02551cfbe

The shark is a right flank. I have done the spot mapping clearly showing first dorsal fin at the right of picture, but it thinks its a left sided flank?

Hi @Sarah_Aus_SpotaShark

I’ve verified this is a bug. We’ll need more time to dig into why you’re seeing right-side matches against a left-side source Encounter. Thanks for your patience.

Weirdly, I tried to re-create the photo and raise it as a separate encounter. I have been adding new sharks to the database without errors, but for some reason it appears that it is this particular image that is making the database assume its a left flank when its right?

That’s interesting. So you were able to re-upload the same image to a new encounter but you’re not seeing the same issues with the wrong viewpoints showing up for match candidates?

I started a new encounter, as if the first image was not there, and did the spot mapping again on that entirely new image, and when I asked the system to find scan, it did the exact same thing and presented the wrong flank options.

It is weird because normally whichever side you start mapping dorsal fins will dictate the flank - e.g. if the first dorsal is on the left then its a left flank. If the first dorsal is on the right of photo then its a right flank - so I am very confused why its producing results on the opposite flank