Document type: article published in La Terre de Chez Nous
Author: Sophie Lachapelle
Preview: Images captured by cameras scrutinizing the slightest movements of cows. This is the basis of a major study currently being carried out by the Innovation Research Chair in Animal Welfare and Artificial Intelligence (WELL-E), jointly created by McGill University and UQAM. Established with $5 million in funding, one of the Chair's objectives is to identify predictors of changes in the mental and physical health of cows. “We want to detect any deviation before it is perceptible to the naked eye, and AI [artificial intelligence] will be able to help us develop early indicators,” explained Elsa Vasseur, Associate Professor in the Department of Animal Science at McGill University and co-holder of the Chair, in an interview on the program La Semaine verte. The ultimate goal? Better profitability for agricultural producers. “There's a misconception that the most productive cows are the most profitable, but that's not the case,” explains Elsa Vasseur's research assistant, Rachel Van Vliet. “These highly productive cows are also sick more often, and that costs money. What's more, a healthier cow will have a longer life expectancy, and that translates into gains.”
The images, in which cows are individually identified, are first analyzed by animal health researchers at McGill University. They determine which behaviors and movements it will be useful to observe. This can be, for example, ear movements, which are very revealing of a cow's emotional state. We can also note very subtle changes in the way a cow walks before she becomes obviously lame. This is important, given that we know lameness to be the number one cause of cows being taken out of service. The analysis is then transmitted to researchers in the UQAM computer science department, so that they can model the data and enable computer learning on a larger scale.
For now, the study, which involves about fifty researchers, is taking place on two experimental farms, one in Quebec (the Macdonald farm attached to McGill University) and the other in Ontario. But in 2025, the University of Montreal will recruit around twenty farm businesses in Quebec to take part in this living lab. “One of the things we’re going to study in the field,” Rachel Van Vliet explains, “is the impact on the cows of moving from tie-stalls to free housing.” The long-term goal is not for every livestock farm to be equipped with cameras to manage its herds. “We want to work with the technologies that they already have, or to identify what the minimum requirement would be”. Dr Van Vliet goes on to explain that the idea of the cameras is to learn, through the videos, about good practices in terms of welfare. "As with humans, we ensure the welfare of cows by meeting their physical and psychological needs. This can include the need for stimulation, curiosity, socialization, etc." The research team plans to analyze the impact of adding certain positive elements to the living environment that could improve animal welfare, such as scratching posts or colored toys for animals. Rachel Van Vliet emphasizes that this is one of the innovative aspects of the research. "We already know that negative factors, such as moving or too much handling, can affect their welfare. Now, we're going to see how we can add potentially positive enrichment and measure the effect that this can have." (...)
