Pig Health Abnormality Detection Based on Behavior Patterns in Activity Periods using Deep Learning

Abnormal detection of pig behaviors in farms is important for monitoring health and welfare. Pigs with problems often have behavioral abnormalities. Observing can help detect take early treatment to prevent disease from spreading. This paper proposes a method using deep learning automatically detecting abnormalities cameras based on behavior patterns comparison activity periods. The approach consists pipeline methods, including individual localization, tracking, abnormality analysis. From measured during the tracking process, healthy pigs different periods day, such as resting, eating, playing periods, were built. Behavioral be detected if behave differently normal same period. experiments showed that built 30-minute time duration over 90% accuracy when applying period-based approach.