Analysis of facial expression in cows as one of animal welfare indicators

Presently, a key component of animal welfare is the animal’s affective (emotional) state and its assessment. Such studies attract great scientific interest focused primarily on negative experiences such as pain, fear, and suffering, which farm animals experience at different stages of their lives. Increased numbers of studies highlight that cows exhibit specific pain behaviours and facial expression as a new and reliable technique which could be developed to recognize and assess pain. Pain in farm animals can be caused by disease, injuries, poor hygiene and housing or inadequate management practice. However, disease (such as mastitis, lameness, peritonitis, etc.) is a major cause of pain in dairy cows, negatively affects welfare, and decreases productivity; therefore, analysis of facial expression can be a valuable early pain detection tool [1–4]. Thus, we aimed to determine parameters of dairy cows (that were affected with subclinical or clinical mastitis) on a facial expresion scale. A total of 30 cows were allocated into equal (N = 10) three groups: 1 (control, healthy cows), 2 (subclinical mastitis) and 3 (clinical mastitis); and photo images (N = 150) based on facial expressions were evaluated. Pain assessment relied on the evaluation of potential pain-related facial expression performances in four regions of the face (each region was scored on a 0–2 scale). Eye and ear position, nostril and facial expression were measured as described in scientific literature [1, 3]. The condition of the changed by 50% (P = 0.07) and 37.50% (P = 0.01), the ear by 42.85% (P = 0.06) and 42.85% (P = 0.04), the nostril by 62.50% (P = 0.18) and 50.00% (P = 0.05), and the facial expression by 33.33% (P = 0.01) and 22.22% (P = 0.001) in groups 2 and 3 of cows, respectively, compared with the group 1. Early detection (changes in a cow’s normal facial expression suggest the presence of pain) of any health problem will ensure that cows can get proper health care as soon as possible, reducing the impact on welfare, productivity and dairy farm economy. We extend our study by developing an automated system (utilizing the power of artificial intelligence) for the detection and analysis of facial expressions.