How a wood tick bite nudged a Ph.D. candidate into computer vision research for dairy cows
Plus, her thoughts on the future of computer vision tech for cows
Just a few months into her Ph.D. program at the University of Wisconsin – Madison, Ariana Negreiro thought she was going to have to drop out. Her Ph.D. research was an intensive feeding trial focused on amino acid and fatty acid interactions. The research required lots of sample collection – including blood, urine and fecal samples. Negreiro considers herself a tough, independent researcher and hard-working student. But the workload was getting the best of her. She just didn’t feel well. Something was wrong with her body.
“I was going to doctors, and they were saying, ‘Oh, you’re just stressed. You’re a grad student. You’re running research. That’s why you’re tired.’ But that went on for a few months until I was at a breaking point because my body physically wouldn’t let me do the things I needed to do. I was exhausted. It got to the point where I had really bad vertigo, and I felt like I couldn’t drive myself anywhere.”
AfiMilk (Sponsor)
Hear how accurate, 24/7 heat detection helped this Wisconsin farm improve preg rates, simplify management and more. See cow monitoring in action.
Negreiro doesn’t come from an agriculture background. The last person to work in agriculture in her family is at least a generation removed. Her hometown is Miami, Florida, and she didn’t fall in love with cows until working as a milker on Michigan State University’s dairy during her freshman year as an undergrad. She found a love for dairy nutrition and that’s what her master’s degree is in.
Faced with the prospect of quitting her Ph.D. dairy nutrition research, she talked to advisers and others at the University of Wisconsin about what to do. It was the spring of 2021 and COVID-19 was still ravaging the U.S. economy. It didn’t seem like the right time to be entering the workforce. She really wanted to continue her education.
One of the advisers she consulted was Dr. Joao Dorea. He and his students were running a trial on the university’s research farm to study machine learning applications for cows. Applied computer vision was the central topic of his research. Negreiro had collaborated some with him already in her research.
Negreiro approached Dorea saying she was looking for other opportunities as she had made the decision she couldn’t continue with the program she had started in.
“I told him, ‘I think your research is super cool research. It seems like it’s the future of agriculture. I just don’t have a background in computer science. I don’t know half of the terms you’re using when we’re talking to each other.’”
Dorea explained that switching emphases would require her take many new computer classes and to learn how to code, but it would allow her to stay on track for a Ph.D.
“The idea of switching at the time was scary for me,” Negreiro says. However, Negreiro decided it was the right fit based on what was happening with her health at the time.
Onfarm Solutions (Sponsor)
Teatwand has been partnering with dealers and farmers for over 15 years, providing award-winning teat spraying technology that reduces labor costs and protects cows from mastitis. Learn more here.
Several months, and many visits to the doctor later, Negreiro was diagnosed at the end of 2021 with Lyme disease. She’s still not sure when or where exactly she contracted the disease, but she acknowledges her love for nature likely put her at higher risk. She loves camping with her dog and being outdoors.
“I always checked myself for ticks after a trip, and I never had the telltale bull’s-eye rash, but you never know where you might encounter a tick. My friend found one crawling on her at the university gym. In Wisconsin, ticks really could be anywhere.”
Negreiro now considers herself a Lyme disease advocate. When she hears others describe similar symptoms to what she experienced – unexplained muscle pain, extreme fatigue, vertigo or shortness of breath – she connects the dots and recommends testing.
“If anyone doesn’t feel right, you should get tested for it,” she says.
In Negreiro’s case, infection from the disease had moved into her heart and brain. Her doctors prescribed long-duration, intensive antibiotic treatments. For a year she received treatment. Even after treatment, she still didn’t start feeling like herself again until just this past summer. While she’s been recovering from Lyme disease, she’s been remaking herself into a cow researcher who is focused on using computer vision.
“I don’t think I would have made such a switch without a big push,” Negreiro says. “I had told myself that I wasn’t good with computers, but it turns out that it’s like anything else – you can learn it if you put your mind to it. I work with Python everyday now, and I’m considered proficient at coding.”
Negreiro’s computer vision research has followed 200 calves from birth through the end of their first lactation. She’s researching how to train cameras to identify individual animals using coat color or key points on their body. She’s using infrared or 2D camera images, as well as depth camera images. The goal is to build an accurate model that can identify animals even as their bodies grow larger and mature. Her works also focuses on application besides identification. For example, a recent project she presented at the American Dairy Science Association’s annual meeting reviewed the ability to predict age of first estrus based off of pre-weaning body shape images to help producers make earlier management decision about which animals to keep and raise as replacements.
She says it’s cliche to say, but in her work, “A picture truly speaks 1,000 words.”
Dorea says Negreiro joined his lab with a background in animal science and has quickly transitioned her training toward data science research.
“Ariana is an exceptionally motivated student, with lots of curiosity and creativity,” Dorea says. “She is advancing her research with great dedication and the highest quality standards.”
She expects to defend her dissertation and graduate in the spring of 2025.
“I’m really excited and happy that this is where life led me,” Negreiro says. “I think there are so many opportunities with computer vision. I think it is the way the whole world is going, and agriculture definitely needs to keep up.”
The Cow Tech Report asked Ariana Negreiro about the current state of computer vision for cows from her perspective and to estimate where the technology is headed in the future. Read more below.