Stuck at 0 individual computer vision tasks complete

Flukebook

Even though the que says its only taking 1 min/ job, and there are only 12 encounters, it hasnt progressed beyond Computer Vision Status: identification requests sent (see table below for links to each matching job). 0 individual computer vision tasks complete of 12 total. Each ID job in the queue is currently averaging a turnaround time of 1 minutes.

note: These are a mix of aerial and regular dorsal shots and I manually annotated all images after a bulk import.

Can you post the link to the bulk import you want me to look at?

https://www.flukebook.org/import.jsp?taskId=a1e379e6-acec-4771-88d2-8e77b98946b4

1 Like

Thanks! I’ll need a little more time to work on this, but I did want to call out that you’ll want to annotate the full body for this species and not just the parts with this visible markings to ensure accurate matches.

Ok good to know! I was just annotating the part where the dorsal fin was. I appreciate your help :slight_smile:

1 Like

Thanks for your patience. We’ve addressed the backlog that caused pending jobs to stall and resent your import to identification. Things appear to be processing normally and this import should complete soon.

1 Like

Thanks Anastasia. I resent after annotating the entire body and while most suggested matches are between other aerial photos within the upload, none are real matches. I am guessing we will need to add more drone photos or try some other perspectives to get the algorithm to recognize IDs and maybe one day link to the existing dorsal IDs.

1 Like