Researchers of Wageningen Livestock Research have been able to define when a turkey takes a step. This is an important milestone to automating gait scoring by sensors. Locomotion obviously is an important indicator of animal welfare. With gait scores based on sensor data, breeding programs are able to better select animals with the best locomotion. This will enhance animal welfare.
Gait scores are determined by trained experts. They examine the animals one by one, which is laborious and time consuming. Sensors have the potential to automatically register a number of characteristics that underlie the experts' gait scores. Sensors have the advantage that they can measure repeatedly and thus follow the scores from day to day.
In this study by the Breed4Food consortium, turkeys from Hendrix Genetics were equipped with advanced accelerometers. The sensors produce information about the turkey's speed and orientation. The researchers were able to determine the start and end times of a step with ‘machine learning’. This step detection model enables fast and accurate step detection in large data sets. Researchers now want to determine the characteristics of each step to predict the turkey's running scores.
These results are published (open access) in a special issue of Frontiers Genetics: High-Throughput Phenotyping in the Genomic Improvement of Livestock. The paper by Bouwman et al. is entitled: Automated Step Detection in Inertial Measurement Unit Data From Turkeys. This work was done within Breed4Food in close collaboration with Hendrix Genetics (Boxmeer) and PDEng students from the Jheronimus Academy of Data Science (Den Bosch).