Optymalizacja YOLOv8 do wykrywania piersi u krów na podstawie danych termowizyjnych

Optimising YOLOv8 for cow udder detection in thermal imaging data to enhance animal health monitoring and farm management efficiency. The study uses a manually annotated dataset with 70% training, 20% validation, and 10% testing splits, created via CVAT. High detection accuracy demonstrates practical potential for automated health assessments in agriculture.