IoT duplicating encounters

In which Wildbook did the issue occur? IoT

What operating system were you using? (eg. MacOS 10.15.3) MacOS 10.15.1

What web browser were you using? (eg. Chrome 79) Firefox 85.0.2

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

What happened? New encounters seem to be duplicating more frequently than before. Sometimes only the duplicate allows for matching process to occur (the original says no matching)

What did you expect to happen? The encounter not to duplicate

What are some steps we could take to reproduce the issue?

If this is a bulk import report, send the spreadsheet to with the email subject line matching your bug report


Could you send me a couple example links? I’m curious if this duplication is a failure to connect the body and head annotations. This would result is one encounter for the body where you could not initiate matching, and a second for the head where you could. This is normal for turtle detection but the encounter should not be split into two records.


Hi Colin

Some of them seem to be creating a duplicate encounter for the body like the examples here (I’ve just linked the duplicated encoutner)

But I’ve also experienced duplications of an encounter that is just a photo of the face. It seems to be creating two boxes around the face slightly offset from each other. No matching can occur on the initial encounter but it can occur on the duplicate encounter. I’ve deleted a lot of the duplicates but here are some examples


Hi, Jill!
I’m planning on diving into this issue Thursday/Friday of this week. I’ll follow up then!
Thanks for you patience, and thanks for providing the examples!

Hi, Jill!

I think that it is as Colin suspected with, meaning that head annotation was split into a separate record.
Unfortunately, all of the other links in the top section that you provided took me to encounter pages that don’t seem to exist any more? I’m guessing that you just cleaned them up manually?

As for the second group where multiple annotations are made for just an image of the head, all three cases made one annotation for the entire hawksbill and a second annotation for just the head, and then split them out. So, I’m thinking that these are actually two instances of the same issue, which is that head annotations are being split into separate records in some cases.

We’ll be tracking this issue internally as WB-1553, and thanks for bringing it to our attention!

Hi Mark

Yes sorry I did clean up those encounters. The issue is still happening though. For example, this clone was just generated, splitting the left side into a duplicate encounter


Hi, Jill!
Yup! It’s a ticket in our system (WB-1553). Either Colin or I will try to address this as soon as possible!

This issue appeared to be linked to the head and body annotations with incorrectly different species detections.

I’ve turned off our automatic assignment for paired annotations and fallen back to the previous heuristic method of pairing detections. This should work well for now until we can retrain the model for better species detection.

Future jobs should not see this error, though there may be a fall in performance for images that do happen to contain multiple animals.

Please let me know if you see this again in the near future. Thanks!

Great, thanks @colin and @MarkF. I’m going to be uploading some new encounters today so I will let you know if we run into the same issue again. We rarely, if ever, upload images of multiple animals so it shouldn’t be a problem in that respect. Meanwhile, I will delete all existing cloned encounters.




Unfortunately, this issue is popping up again frequently now. I am adding my report here instead of opening a new thread.

In which Wildbook did the issue occur?

What operating system were you using? (eg. MacOS 10.15.3)
Various, e.g., MacOS 10.15.7 - issue on all systems

What web browser were you using? (eg. Chrome 79)
Various, e.g., Safari 15.1 - issue with all browsers

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

What happened?
When reporting an encounter, duplicates are created during the annotation detection process, leading to multiple encounters with the same photos or parts thereof. This happens for the following ORP accounts: ORP-Laamu, ORP-North Male, ORP-Lhaviyani, ORP.

What did you expect to happen?
Not this.

What are some steps we could take to reproduce the issue?
Upload a new encounter and wait for the annotation detection to finish. It does not happen all the time, but currently the issue occurs very often.

Links to example duplicates which have not been cleared up manually by users are listed below.

Thanks again for the support!
Best wishes,


North Male

@MarkF @colin - The issue seems to have reoccured, see above post. Please let me know if you would prefer a seperate new thread.

I hope none of the duplicates have been accidentally cleaned up by now, sorry if they have.

Best wishes,