Journal: Data in Brief

Article Title: ActBeCalf: Accelerometer-based multivariate time-series dataset for calf behavior classification

doi: 10.1016/j.dib.2025.111462

Figure Lengend Snippet:

Article Snippet: Data collection , This experiment involved 30 Holstein Friesian and Jersey pre-weaned calves housed in 4 group pens. Each calf was equipped with a 3D accelerometer sensor (AX3, Axivity Ltd, Newcastle, UK; 11g) sampled at 25 Hz and attached to a neck collar from 1 week of birth over 13 weeks. The sensors were removed bi-weekly to recharge the batteries and extract the data. An 8Mp Dome3 CCTV camera was fixed high up in each pen. Behaviors were manually annotated from the videos using the Behavioral Observation Research Interactive Software based on an ethogram (23 behaviors). After timestamp synchronization, observations and accelerometer time-series were aligned using Python (v3.9). No normalization has been applied to the accelerometer time-series..

Techniques: Software