Put two sheep side by side and it might seem impossible to tell them apart.
But new facial recognition technology is proving they do indeed look different.
It could help solve one of the great productivity quandaries for sheep producers — matching ewes with their lambs.
Genesmith managing director Mark Ferguson said his company's facial recognition technology aimed to identify a flock's most productive ewes and, in turn, improve their bloodlines.
"We've got pregnancy scanning so we know who went into that paddock with two foetuses, but we don't know who comes out with two lambs and how heavy those lambs are," he said.
"Some of those ewes might have brought in lambs that are 15 kilos, and another one of those ewes might have two lambs that are 30 kilos each.
"That's a massive difference in profitability."
Dr Ferguson said the technology could detect and identify animals from more than 100 metres away.
He said the camera could scan 360 degrees around a paddock, learning the features of the sheep at a level that would be able to identify them later.
Associate Professor with Sydney University's Livestock and Production Welfare Group, Cameron Clark, said while he had confidence in facial recognition, there were challenges in collecting an original image of the sheep for the camera to recognise.
"In the field, given the environment we've got, how do you get that great image input into the system," he said.
At what cost?
Facial recognition technology is not the only option for farmers to identify their livestock; DNA testing and products such as bluetooth collars are also available.
Dr Ferguson said he hoped his technology could be a cost-effective option.
"One camera or a couple of cameras can handle lots of animals compared to having to put one thing on individual animals, which works beautifully with cattle but the price point is a bit steep with sheep."
CSIRO research scientist Aaron Ingram said despite the costs involved, existing technologies were effective.
"DNA would remain the gold standard as the link between mother and progeny is definite," Dr Ingram said.
"Other approaches have the potential for misclassified relationships."
Dr Ingram said the growing ag tech space would not necessarily make other skills redundant.
"Our goal with a lot of these technologies is to measure things that we haven't been able to measure before," Dr Ingram said.
"It's about adding capacity into the system rather than necessarily replacing capacity."
Dr Ferguson said the trial of the technology would nvestigate many of the challenges being raised.
"We have no interest in making a technology that doesn't serve farmers," he said.
"The idea is to make sure it's practical and that the technology really will change the game."
