Hi Jason, 
I've had Ibeis and Wildbook on my radar for ages and have been meaning to ask you about the two projects, so this is a neat opportunity! Could you tell us a bit more about the two projects? I have been trying to figure out how they are integrated (and also seperate - it's a bit confusing from the outside!), so I'm interested to find out more about where you hoping to take the project(s).   

Is the platform open to anyone interested in analysing images for patterns? How does it work? 

Cheers,

Steph 

Hi Steph,

I would say so far at Perhentian Turtle Project, we have not found the shell damage as an issue for identification. 

Primarily, we use facial scale patterns for identification. In addition to that, some individuals have unique natural markings or scars (on top of head, shell, flippers) which then are used as secondary identification.

Personally, I found that these markings/scars is actually somewhat benefical in terms of identifying individuals. When the photographs were unsuitable to process by pattern recognition software due to low quality, we manually identify individuals through visual matching using identification keys and these markings/scars is one of the key features we are looking at. 

Also, we have found as well that the damages/scars on top of head/shell/flippers helped us in monitoring the individuals. For example, we have been observing one female turtle since 2012 and in 2015, we found her with a cracked shell and believed she was hit by boat propeller. Until now, she survived the hit and is still feeding at the foraging ground. From this, we can actually monitor her in-water recovery progress.

Regards,

Nazirul

Hi Steph,

Sorry for the late reply as I have other commitments currently. 

In our case, we are using facial scales pattern because we found out that each individual have a distinctive scutes pattern (number, position & shape) on their face. Fins pattern are usable as well for identification but due to its complexity, the chances for errors (misidentification) are greater. This however can be solved when we have a more species specific pattern recognition software which can pin-point the exact distinctive features on the fins pattern.  

Based on our findings in 2015, we found that facial and fins patterns are more reliable than the shell markings as it provided more details in identifying individuals. There are some individuals with very distinctive shell markings but unfortunetly, it does not apply to the whole population especially the juveniles. Most juveniles have almost similar shell markings. 

In Perhentian Islands, observation of sea turtles while snorkeling or diving has become an opportunity for us to conduct the study. We have a team of trained research intern who went out for snorkel survey looking out for turtles every day throughout the season. Whenever a turtle is sighted, one of the research intern will skin-dive and photograph the top view of the turtle and both sides of the face. All the photos are then brought back to the research station and analyse using the pattern recognition software. The reason why we took the top view photo is because it is easier for us to know the sex of the individual when we analysed the photos.

Sorry to tell you that our website is currently not available. We have been trying to bring it back but there is a constant virus & spyware attack. We are now reachable only by Facebook page:  https://www.facebook.com/perhentianturtleproject/?fref=ts

Regards,

Nazirul