Study with USDA researchers affirms startup’s AI-powered facial recognition for cows can detect sick animals

A collaborative study between an Overland Park animal agtech startup and the U.S. Department of Agriculture successfully corroborated MyAnIML’s predictive ability to proactively manage devastating disease outbreaks in cattle production, the company announced this week.

Shekhar Gupta, MyAnIML; photo by Taylor Wilmore, Startland News

MyAnIML uses first-of-its-kind, proprietary facial recognition and deep learning technology to analyze cow muzzles — accurately predicting Infectious Bovine Keratoconjunctivitis (IBK), or bovine pinkeye, 99.4 percent of the time and several days before veterinarians were able to detect symptoms, according to published study results.

RELATED: How a KC startup is using Bluetooth to help ranchers ID sick cows days before symptoms

“Early detection of disease is critical to healthy herd management — giving producers the chance to separate sick animals, control spread and judiciously use antibiotics before a large outbreak occurs,” said Mike Clawson, a USDA Agricultural Research Service molecular biologist and project researcher who alongside scientist Larry Kuehn has studied IBK for years at the U.S. Meat Animal Research Center. 

“The results of the MyAnIML study demonstrate how far and how impactful AI-powered technology can be toward ensuring a safe, resilient and sustainable U.S. food supply chain,” he continued.

The MyAnIML and USDA study included 870 beef cattle located on three different Kansas ranches during the summers of 2021 and 2022.

Bovine pinkeye is highly contagious and the most common ocular disease of cattle globally, costing U.S. producers alone an estimated $150 million annually in lost performance and treatment costs, according to MyAnIML. With no effective vaccines for IBK, producers must treat infected animals with antibiotics, thereby elevating the risk of developing antibiotic-resistance bacteria strains that threaten human health.

Building off this collaboration, MyAnIML and USDA are applying the technology next to predict Bovine Respiratory Disease (BRD), said Shekhar Gupta, CEO and founder of MyAnIML. As the single most economically impactful cattle disease, BRD costs the U.S. feedlot industry potentially more than $900 million annually, according to the American Society of Animal Science.

“Building off our AI platform’s ability to distinguish between ‘healthy’ and ‘sick’ cattle two to three days before symptoms were diagnosed, the next step is expanding MyAnIML’s repertoire of diseases and health events to offer producers new tools to ensure a safe food supply,” Gupta said.

MyAnIML uses images of cow muzzles to predict impending illness days before symptoms occur. This photo was taken by a GoPro camera mounted on a feed truck and uploaded to the MyAnIML AI-driven technology platform for analysis; photo courtesy of MyAnIML

MyAnIML’s patent-pending technology platform uses AI, facial recognition and inexpensive GoPro cameras to automatically capture and analyze subtle changes in a cow’s muzzle.

Like a human fingerprint, each cattle muzzle is unique, and can be used to track specific cows. However, MyAnIML, is the first to use muzzle dermatoglyphics to not only identify cows, but as a health predictor of diseases such as BRD, IBK, foot root and uterine infections, proving that subtle changes in the bumps and ridges on a cow muzzle are a precursor of a health event, according to MyAnIML.

MyAnIML’s trial partners have also successfully used the technology to monitor the health status of cows going into estrus, early stages of labor and even subtle health stressors, like the need for more nutrition while nursing calves, Gupta said.

“The muzzle is an incredible mirror into cattle health and well-being. Humans just didn’t have the ability to ‘see’ what the muzzle was telling us without the help of MyAnIML’s advanced AI and facial recognition technology,” he said.

Since its founding in 2021, MyAniML has:

• Developed a proprietary dataset of 3000 muzzle images of beef cattle, the first cattle facial and muzzle image library for health management.
• Partnered in trial projects with multiple Midwest cattle production facilities.
• Released an inexpensive Bluetooth-enabled “smart” ear tag and app that helps large-scale commercial feedlots and stockyards quickly locate potentially sick animals identified by the MyAnIML predictive platform.
• Signed a partnership agreement with DairyFi, an India-based startup, in January. With 308 million cows, India has the largest cattle herd in the world, but suffers from lack of adequate cattle health services, greatly reducing the economic and productivity potential of India’s cattle industry.

With positive results from the collaboration, large-scale trials, veterinarian feedback and ongoing product development, MyAnIML is getting ready for broad-scale commercialization, Gupta said. 

The company received angel investment funding in 2021 and is currently in the process of a seed funding investment round to expand its technology in cattle, as well as to include other livestock and companion animals.

Click here to read more about the startup journey of MyAnIML, one of Startland News’ Kansas City Startups to Watch in 2022.