Wildbook - AI to track individual animals from images

Sharing in case of interest: http://wildbook.org/doku.php

Does anyone here work on Wildbook or know more about it? Here’s how the effort is described at the link: “Wildbook blends structured wildlife research with artificial intelligence, citizen science, and computer vision to speed population analysis and develop new insights to help fight extinction.”

Some of that overlaps with iNaturalist, but Wildbook’s added value to our community seems to be the potential around AI to detect and track “individual animals in a wildlife population using natural markings.” iNat and Wildbook share some of the same funding sources I think, and contribute to the some of the same repositories like GBIF. Any collaboration happening yet between iNat and Wildbook?


This sound like it could be tremendously useful and interesting if it works well. I sure hope it does. Thank you for bringing it to our attention.


Thanks for sharing, this sounds like an amazing project and the idea of combining the iNat image base with image classification algorithms developed by Microsoft or by the researchers involved in Wildbook sounds like an absolute no-brainer to me.

1 Like

You’re welcome, glad you and others found it interesting. To clarify, I’m content (and frequently awed) by iNat’s in-house image classification system in terms of taxonomic suggestions – I don’t have any reason to envy what Microsoft et al can do. It does seem like for the species where researchers are advertised as using Wildbook…

  • Giraffe
  • Whale sharks
  • Manta rays
  • Sea turtles
  • Leopard sharks, ragged tooth nurse sharks, seven gill sharks
  • Jaguars
  • Cetacean (flukes)
  • Grevy’s zebras
  • Polar bears
  • Iberian lynx

… it seems like generally a good thing for Wildbook to either harvest images directly from iNaturalist or via GBIF to identify and track individuals. What I would personally like to see in response is feedback/engagement by Wildbook on my observations. Like, “hey, that’s individual #547b, last seen a year ago and 10 km away” or “that’s a new individual we didn’t have in our records, thanks!” Relatedly, I’m a sucker for learning the individual back-story of birds that I observe and report banding data for, and used to work for an African wild dog project that manually tracked hundreds of individuals via photos, so this kind of effort is of particular interest to me.

I would also generally like to see more cross-initiative/cross-platform collaboration where data is harvested from iNat and I’m not asked to submit my observation in multiple places.


I’m interested to hear from others on this and where the boundary is for machine-generated content/updates like this. i.e. it wouldn’t be a person giving you that information but would come automatically via an integration. Do you see any downsides to this kind of interaction?

I’ve personally intersected with a few different members of the WildMe team over the years but there hasn’t been a serious conversation about integration. I’m long overdue to check in with them on if/how they’re using iNat data (I’d be surprised if they aren’t) and explore what kind of collaboration might be mutually beneficial.


This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.