discussion / Marine Conservation  / 2 July 2016

Sea turtle identification through pattern matching

This week I have been attending the Association for Tropical Biology and Conservation Conference in Singapore. Although marine conservation is not the main objective of the conference there have been some fellow marine conservation scientists here doing some interesting things.

The Large Marine Vertebrates Project LAMAVE and the Perhentian Turtle Project, from the Philippines and Malaysia respectively, have been using pattern recognition software to help identify sea turtles to the individual level. Pattern recognition software is more commonly known for whale shark ID and sometimes sharks in the marine environment but for turtles this is relatively new.

It would be great to hear if there are other projects or people using this kind of software. At first glance it seems like this method could be far less invasive and cost-effective than other mark-recapture methods, like tagging, so it would be great to hear if there are more examples of its application.

Also are there other methods out there or that are being developed that could be used for mark-recapture that don't require tagging?!

 




I think at least the Perhentian Project is starting to use the Wildbook software platform (data management and image analysis). Information is currently spread across two sites: wildbook.org and ibeis.org...projects merging. It's an exciting field with a lot of opportunity to be less invasive but also much more cost effective in terms of managing larger volumes of data.

Hi

If anyone has any questions about our studies in the Perhentian Islands, give me a shout :-)

 

Dan

Hi Daniel, 

We had a comment come back on twitter asking whether shell damage from mating, reef & oat strikes, and barnacle growth may impact it's effectiveness - have you found this to be an issue? 

Cheers, 

Steph

Hi Nazirul, 

That is so interesting! My assumption was that you would be using the shell markings as the ID, I hadn't even considered that the facial and fin markings would be a more useful way to diferentiate individuals. I would have thought that getting a look at a turtle's shell would be easier than taking a photo that shows the fins or face in enough detail to analyse - is this not the case? Or do you use the facial scale patterns because the shells do not differ enough between individuals?

I'm also keen to hear more now about your methodology. How to you take your observations? If you're using pattern recognition software and Wildbook, you must be taking photos. Are these remotely triggered (i.e. underwater camera traps), or are they from photos taken by divers or people there monitoring in person? 

Kate's link above doesn't seem to be working for me, is there a website we can visit to find out more? 

Thanks in advance!

Cheers,

Steph