Luara A. Freitas, a postdoctoral researcher in Guilherme Rosa’s lab in the University of Wisconsin–Madison Department of Animal and Dairy Sciences, won first place at the 2026 UW Postdoctoral Research Symposium for her research on identifying dairy cows using multiple camera views and machine learning.
Her poster, “Multi-View Fusion and Cooperative Learning for Individual Identification of Dairy Cows Under Occlusion,” explores how combining information from different camera angles can help identify individual cows even when other animals or barn structures partially block the view. The study’s best-performing approach achieved 94% identification accuracy, demonstrating the potential of this technology for automated livestock monitoring.
The research is part of a broader project developing ways to identify and continuously track individual cows in barns using low-cost surveillance cameras. This monitoring could help farmers measure feeding and drinking behavior and detect changes that may signal illness or metabolic problems, supporting timely management decisions.
Co-authors include Rafael E. P. Ferreira, João R. R. Dórea, Kent A. Weigel, and Guilherme J. M. Rosa. Ferreira, formerly a doctoral student and postdoctoral researcher with Dórea, is now a scientist at UW–Madison’s Data Science Institute.