AIHerd’s solution is designed to help farmers understand the reasons for their animals’ behavior, quickly alerting them to unusual behavior (e.g. heat or disease) in an animal. An individual behavior “signature”—robust over long periods of time—is calculated using CEA-List’s AI and vision algorithms.
Machine learning without human annotation is one of the major innovations to result from this project, and it is what made it possible to very quickly get actionable information out of the huge amounts of data acquired by AIHerd [1]. This AI model delivers 80% accuracy detecting animals in heat, a success rate far superior to competing wearable technologies.
[1] Cumulative unsupervised multi-domain adaptation for Holstein cattle re-identification. Fabian Dubourvieux, Guillaume Lapouge, Angélique Loesch, Bertrand Luvison, Romaric Audigier. Artificial intelligence in Agriculture, December 1st, 2023.
Correct heat detection rate
are using the solution
This project is the perfect illustration of the CEA’s capacity to transform research results into technologies that can respond to the most demanding use cases.
Our motto is to bring improved wellness to livestock and financial benefits to farmers.
