Aussie researchers to test koala 'facial recognition'

BRISBANE — Australian researchers are launching a koala "facial recognition" study that aims to better monitor the marsupials' behaviour and ultimately aid in their conservation.

The team from Griffith University is attempting to use artificial intelligence (AI) to recognise individual furry mammals when they use wildlife road crossings in Queensland state.

The underground pathways and bridges are installed at busy roads near koala habitats to offer them a safer route away from cars. 

Associate professor Jun Zhou, who is leading the pilot study, said it was hoped AI would end the need for manual camera checks to determine which species were using the crossings.

"Now, with artificial intelligence developing very quickly over the past 10 years, the technology is powerful enough to help recognise not only koalas generally, but which individual koalas are using the crossings," he said.

Researchers have previously also used identification tags and GPS to track koalas.

Griffith University said the team would work with conservation groups on training AI to distinguish between individual animals based on their appearance and movements.

They hope the data will offer a greater understanding of how koalas use the wildlife crossings and if the crossings could be better located to protect the animals from getting injured by cars.

The government-funded project is set to get underway at 20 camera sites near Brisbane in July.

Koalas live along Australia's eastern coast, but are considered vulnerable to extinction and face a raft of threats including habitat loss from logging, development and bushfires which destroyed swathes of forests in 2019-2020. AFP