Dear Internet of Turtles Community,
On Monday, February 2nd, we will be deploying two updates to the Internet of Turtles Wildbook:
1. Cross-Side Viewpoint Matching for Hawksbill and Green Sea Turtles
Based on emerging research in AI for sea turtle individual ID, we will be supporting broader matching against different viewpoints, especially for matching left and right head scutes. Currently, left-side photos are only matched against other left-side photos (and right to right). We will be relaxing this constraint to allow left-side queries to match against right-side candidates and vice versa.
This change is supported by recent research demonstrating that deep learning models can detect inherent visual similarities between an individual sea turtle’s left and right facial profiles.
Reference:
Adam, L., Papafitsoros, K., Jean, C., Rees, A.F., & Čermák, V. (2025). Exploiting facial side similarities to improve AI-driven sea turtle photo-identification systems. Ecological Informatics, 89, 103158. https://doi.org/10.1016/j.ecoinf.2025.103158
These authors note: “We show that the similarity between the left and right profiles of the same individual with respect to geometry, coloration and pigmentation, is on average higher than the similarity between profiles of different individuals. The similarity is detectable even when images are taken years apart and under diverse settings and conditions. We perform several image retrieval experiments under scenarios which mimic realistic sea turtle photo-ID matching processes, where we also allow comparisons of opposite sides in the matching process which have no spatial overlap. We show that the detection and exploitation of this similarity is translated to improved accuracies when compared to the traditional side-specific image retrieval setting.”
Based on these findings, we believe that supporting cross-viewpoint matching brings the IoT community in line with the latest research in AI for sea turtle re-ID. However, we know that IoT is also used in real-world photo ID, and we value your initial thoughts on this change and your feedback in using it.
2. MiewID v4
We’re upgrading the base AI matching model from v3 to v4, which provides improved accuracy for individual identification. Notably, MiewID v4 is trained with left and right viewpoints combined under the same sea turtles individual IDs, making it inherently capable of detecting cross-side similarities, such as those suggested in the paper above by Adam et al. 2025.
What you may notice:
- More candidate matches returned per identification task
- Some matches may show opposite-side photos of the same individual
As always, please let us know if you have questions or observe any unexpected behavior after the update.
If you believe now or in the future that cross-viewpoint matching does NOT help with your use of IoT, please let us know via email or on community.wildme.org. We’re happy to turn it back off again based on community feedback.
We’re a Global Community
Wildbook is now a global community of over 2500 users tracking over 307,000 individual animals across 294 species.
We welcome your feedback, support requests, and any other questions at https://community.wildme.org or via email at wildbook.community@conservationxlabs.org.
Thank you for letting us support you in your work!
Dare Mighty Things,
The Wild Me Team at Conservation X Labs