Our next guest is Eileen Roy-Zokan. Thank you for being here and welcome to The Junction. It's my pleasure.

Thanks for having me on.

Can you tell us about your background and how it relates to the field of veterinary forensics?

So I have a doctorate in genetics and I've been working in genetics for 20 years, so I mostly do DNA analyses. Most of my cases deal with wildlife versus domestic animals, so I'm usually working on poaching cases and um, species identification, things like that.

I just actually, uh, finished a class on international wildlife trafficking and, and law. Oh, very cool. Yeah. So, I was, I was, I was learning so much about the different laws and, And also, like, the things that people are doing, because it's so much complex that I even thought.

Right? Oh, so much. The networks are, are crazy when it comes to trafficking of wildlife.

Yeah, and the excuses they use and the type of species that they take and the reasons for it. So, I mean, it is just so complex.

It really is.

So, what was your topic today?

My topic was something really outside of my field. It was on photogrammetry. So, um, photogrammetry is the science of getting information about an object or the environment that's featured in photographs. And so my talk was kind of focused around that.

So I love taking pictures and I love manipulating pictures and using things like Photoshop. And I was really excited to see that it's something kind of similar. So how does, how does photogrammetry work?

Traditionally, in a normal setting, in a normal crime scene, um, you're going to be using, you're going to be taking a bunch of overlaying pictures of your crime scene, and you can upload it on special photogrammetry software, where then it will recreate, do a three dimensional reconstruction of your crime scene. Um, They also use this a lot in archeological digs to kind of map out their dig area and also ancient artifacts that kind of have a digital copy of that artifact that can be safely examined afterwards. In my case, I can't do that because the photographs I'm using are, are. Stuff like, um, social media posts and, um, game camera pictures. And with these photogrammetry software, you need, the camera that you're using needs to be known by the program, meaning you've taken it through a series of calibration tests, that the program then is able to, any of the pictures that you take with that camera, is able to extrapolate the depth of field. from each of those photographs based on the calibration tests. And then that's how it's able to get that geometric information from your photographs to build that three dimensional model. When you don't know the camera, like you're taking a picture off of Facebook, that's not possible, so I can't use those traditional softwares to do my analyses.

And I imagine also the fact that you're dealing usually with probably like a single shot versus multiple shots, that's

Absolutely. That is absolutely the case. Um, I'm usually dealing with just one or two pictures, um, that I have to compare. And there's basically three different analyses I do. One is individual identification. I'll get a couple pictures and they're like, is this deer in this social media post the same deer that's in, on this game camera picture? Or I'll get two photographs of like, I had a photograph of a, of a deer mount that was taken in a residence about, you know, two decades ago versus a deer mount, a picture of a deer mount that was taken. And another individual's residence two years ago, it was some arguments between neighbors, one stole something from the other. And it's just, is that the same deer Mount? The other thing I'll have is comparing social media posts or game camera pictures to a physical object. We have an officer seize, you know, an antler rack. And they have some game camera pictures. Is that, did that rack come from that deer that they have on the game camera or the social media posts?

One of the examples that you posted was as far as like doing, taking measurements of a deer antler. And that was, if I understand it correctly, to see if it was like poached the correct time or, or taken at the correct time.

Yeah. So here in the state of Florida, we have a regulation that, uh, male deer cannot be taken if their antlers are less than five inches. So there was a social media post of an immature buck that the person that posted the social media post, um, they actually strapped on an eight point rack on On the, on the immature buck using a towel to make it look bigger than it actually was. Um, and our officer seized the rack and they had two questions. Whether the rack they seized was actually the antler rack that's featured in the picture. And then in the picture you could also see the actual antler point of the immature buck. Could I somehow get an estimate of the size and length of that antler point? And because, um, What I have to use is either a program like Photoshop or another free version of Photoshop known as GIMP. to try and get the number of pixels along these objects. I so you need an object of known size in the same plane as the object you're interested in measuring. And so in this picture, I saw that one of the antler points on the rack that was used as a toupee was right next to that. actual antler point of the immature buck. And so I used the rack that I had in the lab, um, that was seized. That was the actual rack in the used as a toupee. Um, I measured that out as about 2. 6 inches. And then I was able to then look at the, I got the number of pixels. Um, so I measured out, the length on GIMP and got the number of pixels that was contained along that particular antler point on the toupee rack and then I got the number of pixels of the actual antler point on the immature buck and then from there what I do is I know the size of the antler point from the physical rack I have in the lab it was about 2. 6 inches. And so what I'm able to do is say, okay, this number of pixels for that particular antler point divided by its actual size is this number of pixels per inches. So how can you use that as a conversion factor, because you know your number of pixels per inch. And then from there, I was able to measure the number of pixels along the antler point on the immature buck, divide that by the conversion factor, that's your number of pixels per inch, and that will convert that to, will give you the size estimate for the object you're interested in. Now this hasn't been court validated. We are testing it in the lab to get the statistics and variations behind these kinds of estimates, but that's kind of the approach that I use.

So this is a relatively early field, right? So there's still a lot. So where do you see this going as we move forward in forensics?

Oh boy. It would be really great if we can kind of automize this a little bit. Um, we also need to test out different programs. I know there was a dissertation recently that was published back in 2019 where they, Did these kind of analyses, but they use the Pharaoh scanner as well and Photoshop and Gimp and compared all three methods and although their sample sizes were low, it's very promising. So there's so much technology development that can happen, but if they could get that more automized and the statistical variations down, so it's going to take a lot of testing of different environmental factors, different angles and things like that to get that more up to date.

Right, well I'm looking forward to seeing where this takes us because I think that there's a lot of information with how many people take pictures of everything and so many sources. I think that this is going to be really useful and relevant in crime scene investigations, especially with animal crime scenes and as you mentioned with like wildlife, uh, illegal wildlife trade. Absolutely. Yeah. Yeah. So, thank you so much for being here and thank you for talking to us and thank you for what you're doing.

It was

my pleasure.