
Photo by Photo Submitted
CHEESE – The research student is beginning her thesis work with a focus on animals that do not lose their horns every year, such as this mountain goat. Photo by ISOBEL NESS
Photo by Photo Submitted
CHEESE – The research student is beginning her thesis work with a focus on animals that do not lose their horns every year, such as this mountain goat. Photo by ISOBEL NESS
Photo by Photo Submitted
TELLING THE DIFFERENCE – As her thesis for her masters, local post-graduate student Isobel Ness is hoping to advance facial recognition technology to positively identify individual animals from photographs. Photo by ISOBEL NESS
Photo by Photo Submitted
TELLING THE DIFFERENCE – As her thesis for her masters, local post-graduate student Isobel Ness is hoping to advance facial recognition technology to positively identify individual animals from photographs. Photo by ISOBEL NESS
Photo by Photo Submitted
PATIENCE REQUIRED – Ness says it can take a while before she’s able to get the shot she needs. Since beginning her research work this past fall, she has taken hundreds and hundreds of photos at the Yukon Wildlife Preserve. Photo by ISOBEL NESS
Photo by Photo Submitted
Cartoon by AMANDA GRAHAM
Photo by Photo Submitted
ISOBEL NESS
A local post-graduate student at Yukon College is focused on advancing methods to recognize wildlife by facial characteristics and other features.
A local post-graduate student at Yukon College is focused on advancing methods to recognize wildlife by facial characteristics and other features.
Isobel Ness is conducting the research as part of her thesis for her master’s in northern environmental and conservation sciences.
She took the four-year degree course offered at the college in partnership with the University of Alberta.
Being able to positively identify an animal through specific features – whether it’s the distance between its eyes, length of its ears – would have a number of practical applications in wildlife research and management, Ness explained in a recent interview.
She acknowledges physical features of different animals already help scientists identify individuals, such as using the fluke of a whale’s tail or distinct markings on orcas.
Ness wants to first develop a method to identify ungulates – animals with hooves – who don’t lose their horns each year and animals who shed their antlers annually.
The second step is to take the method into the field and apply it to the regular practice of using camera traps, or cameras triggered by movement that are set up in remote locations to monitor wildlife, she says.
“Obviously, all animals have something,” Ness says of distinguishing characteristics.
“You just have to find it, figure it out.... It’s taking longer than I thought because trying get the animals to look at you head on is more difficult than I imagined.”
Ness says having a method to recognize animals by facial features may be able in some circumstances to displace aerial surveys, which can be disruptive for animals.
It could be used to investigate complaints of problem wildlife to determine if it’s the same individual getting into the hay shed or what have you.
Confirming the identification of individual animals in the field could provide information needed to help figure out population densities and other demographics, she says.
Ness says the technology might be able to provide some information normally gathered through radio collaring which requires tranquilizing the animal – usually from the air – to attach the collar.
Senior wildlife biologist Tom Jung of Environment Yukon says when Ness was looking for ideas for her thesis, she thought facial recognition would be of interest to Environment Yukon.
“So we helped her out a bit,” Jung said in a recent interview. “We are always looking at ways to do things better.”
Jung suggests identifying a tiger by its distinct stripes is one thing. Being able to single out a deer from 100 photos is another, he says.
“How do you know they’re not the same two or three?
“Being able to single out individuals in wildlife populations is really big,” says the senior biologist. “It’s really a challenge to do with most of the wildlife in the Yukon.”
Jung says there are a lot of tried and true methods to count various species, various ungulates.
There are ways to count caribou, to count bison, but they’re always searching for methods which are less expensive and less invasive, he says.
Jung says there are ways to track individual animals but it usually means catching the animal first, at significant expense.
“It does not mean we are not going to be invasive, that we can get around it 100 per cent,” he says. “But in cases where we think we can, it is incumbent upon us to try.”
Jung says by camera-tracking individuals in the wild, biologists might be able to extrapolate local population densities, for instance.
Since beginning her research this fall, Ness has taken hundreds and hundreds of photographs of the captive populations at the Yukon Wildlife Preserve.
Staff at the preserve have been quite accommodating and supportive, she says.
Ness says the wildlife, on the other hand, don’t always do what she’d like them to do, including looking at her when she’s ready to snap a picture, or holding still for a group shot before she zooms in on the individuals.
She’s concentrating first on the bovids, the animals that don’t lose their horns every year, such as the mountain sheep, goats and muskox.
Because the muskox population at the wildlife preserve might not be large enough to provide the diversity she needs, Ness says she may have to head up to the Large Animal Research Station managed by the University of Alaska Fairbanks to augment her library of muskox photos.
Ness says she’ll then turn her focus to the cervids, the animals that lose their antlers every year, such as the deer, elk and moose.
She expects her thesis could take 2 1/2 years to complete.
“It might take a little longer, but we’ll see.”
Ness says she’s hoping this winter to create enough of a foundation in her research that she’ll be able to begin setting up camera traps in the spring.
In order to encourage thoughtful and responsible discussion, website comments will not be visible until a moderator approves them. Please add comments judiciously and refrain from maligning any individual or institution. Read about our user comment and privacy policies.
Your name and email address are required before your comment is posted. Otherwise, your comment will not be posted.
Comments (4)
Up 9 Down 12
ralpH on Dec 27, 2016 at 11:45 am
Do we really need to spend this kind of money on this. Seems Health and Education could make a case based on need. Besides how they taste is all that matters. lol
Up 9 Down 2
Alex Gandler on Dec 26, 2016 at 7:42 pm
I LOVE the goat in the picture - he looks like he knows so many things that
we never will.
Up 14 Down 10
Just Say'in on Dec 25, 2016 at 1:22 am
How about we start using this technology to put some criminals behind bars and just leave the animals alone and quit harassing them.
Up 10 Down 2
ProScience Greenie on Dec 23, 2016 at 3:44 pm
Great news. Hope the tech works out even better than expected. It sure would be nice to see an end to the barbaric practice of tranquilizing and collaring wild animals.