MANHATTAN, Kan. (KSNW) – It’s been almost two years since K-State professor KC Olson approached the Kansas City-based company Black Hereford Holdings, LLC about a bovine facial recognition app.

“I hadn’t really thought about an application in the livestock industry, but the idea was really intriguing,” Olson said.

Olson and a team of K-State Research and Extension staff members provided a proof of concept for the app in 2019.

“I was fortunate to gather some really talented colleagues from the Animal Sciences, from Veterinary Medicine and from our Computer Science Department,” Olson said.

Olson says initially, the team believed they needed to approach the app the same way they would approach developing human facial recognition.

“We took this to our colleagues in computer science, and they kind of smiled, and they said, ‘well, we’ve really advanced beyond that now, and we think we can do this very quickly with a form of AI rather than a mathematical modeling sort of approach to the head,'” Olson said.

Months after the proof of concept was finalized, Black Hereford Holdings launched CattleTracs in January 2021, the app utilizing artificial intelligence to learn and ‘memorize’ individual cows’ faces.

“The only thing that transits with the photographs are the GPS coordinate where the photograph was taken, and the date,” Olson said.

Scarlett Hagins with the Kansas Livestock Association says those coordinates can be used for contact tracing, a crucial tool for the state’s top agricultural industry.

“We’re talking 6.5 million head of cattle in the states, that’s twice as many people in the state. So, any kind of technology that would allow us to do that just like this facial recognition technology would be very important to farmers and ranchers,” Hagins said.

Olson says as the database of cattle grows, the app will be able to track the entire lifespan of a cow in the system.

“Their age, their origin, any kind of health procedures they’ve had performed, and whether or not they’re eligible for things like natural or organic or value-added markets,” Olson said.

Olson says another future update includes a high-speed photographic platform more suitable for use where there are larger concentrations of animals.

“It gets a little cumbersome, you know, for a feed lot that accepts truckloads and truckloads of new cattle every day, so we’re working with another company right now to develop that, that high-speed shock-proof automated photographic platform,” Olson said.

Olson says while 70,000 head of cattle have been added to the database, it’s not nearly enough to warrant future updates.

“We need a half a million for the AI to be able to mature and to maximize predictability and accuracy,” Olson said.

For more information, visit the K-State Research and Extension website here.