AI-powered acoustic surveillance for early detection of calf respiratory disease

[Federica Motterle]

Effective management of Bovine Respiratory Disease Complex (BRDC) in calves requires early, non-invasive diagnostic tools. This study investigated the potential of an AI-driven acoustic monitoring system to detect coughing, a primary early symptom of BRDC. Researchers recorded over 2,730 hours of audio across 30 days in four pens housing seven calves each. Using a lightweight HuBERT-based AI model, the system achieved a 92% accuracy rate in classifying cough events. The temporal patterns of these coughs closely mirrored infection dynamics and treatment responses. Crucially, the AI system detected increased coughing one to two days before traditional clinical symptoms appeared. These results demonstrate that autonomous acoustic surveillance is a highly sensitive, scalable, and valid tool for the continuous monitoring and early detection of respiratory diseases in livestock. Read the full article on Animal.​