Equine Welfare Assessment: Horse Motion Evaluation and Comparison to Manual Pain Measurements
Pain estimation for horses is a tedious task since they are flight animals and tend to suppress pain specific behaviour. We obtained continuous video data of horse in different levels of pain, to evaluate their behaviour – while they were undergoing routine castration. During the whole time we regularly and manually evaluated the horses’ pain.
To quantify the horses’ motions, we automatically extracted horse masks from which we derive their orientation and position. We then performed a motion feature selection based on the different types of manual pain measurements. This is thus a first time comparison of long term manual and automatically derived equine pain evaluation.
A result is the decreased motion entropy of horses in pain and a tendency of staying in a place for longer periods of time. This was reflected with large observation time windows and features related to this behaviour – features which measure the entropy, for instance, or the change quantiles. It also turned out horses behave differently when humans are nearby, thus this method gives an unbiased view on the pain behaviour, especially motion under pain conditions.
These results can decrease the workload in veterinary clinics and provides means of remotely displaying potential discomfort of horses.
KeywordsImage processing Signal processing Equine welfare monitoring Time series Pain estimation
- 3.Chen, L.C., Papandreou, G., Schroff, F., Adam, H.: Rethinking Atrous Convolution for Semantic Image Segmentation. arXiv (2017)Google Scholar
- 9.Merskey, H.: Pain terms: a list with definitions and notes on usage. Recommended IASP Subcommittee on Taxonomy. Pain 6(3), 249 (1979)Google Scholar
- 10.Nasirahmadi, A.: Development of automated computer vision systems for investigation of livestock behaviours (2017)Google Scholar
- 14.Reulke, R., Rueß, D., Deckers, N., Barnewitz, D., Wieckert, A., Kienapfel, K: Analysis of motion patterns for pain estimation of horses. In: 15th IEEE International Conference on Advanced Video and Signal Based Surveillance, pp. 1–6 (2018)Google Scholar