ERA - SOUNDWEL - Entwicklung eines Werkzeugs zur automatischen Echtzeit-Ermittlung von Emotionen und Wohlbefinden in den Lautäußerungen von Mastschweinen

Contact: Dr. Sandra Düpjan

Duration: 2016-2019

Funding: Bundesministerium für Ernährung und Landwirtschaft, Bundesanstalt für Landwirtschaft und Ernährung

Recently, the importance of the impact of mental health on animal welfare has become evident, and animal welfare is now assessed through both physical and mental health. However, tools to scientifically and easily measure the affective lives of animals are still lacking. One promising tool to assess animal emotions is through vocalizations. Our project proposes to gather several teams that have conducted research on vocal indicators of mental states in pigs, in order to develop a robust system for identifying emotions in fattening pigs (from birth to slaughter), using vocalisations, which could be used to obtain welfare outcome indicators on-farm. The project includes three main tasks. First, the audio data owned by the teams will be gathered and sorted according to the emotions inferred from the recording situations. These data consist of pig calls recorded in situations ranging from stressful husbandry procedures (e.g. castration and handling) to positive situations (e.g. social reunion and post-nursing interactions). Because the recordings vary in terms of pig breed, sex, age and recording technics, we will also carry out more recordings, using the same situations, but in standardized conditions, in order to obtain a baseline for the recognition system. Additionally, to cover a broad range of situations encountered by pigs during their lives, important husbandry-related situations that are missing from the database will be added. This first part will end up with a meta-analysis of the gathered data, aiming at identifying vocal indicators of important emotion-related characteristics of the recording situations (e.g. emotional arousal and valence). The second task will consist in testing different classification methods/models, in order to choose the more appropriated to develop a system that could reliably identify pig emotions. The third and last task will consist in testing and validating the system. To this aim, the teams will test the system in various situations at their own research stations and on-farm. The final outcome of this project will be an accurate non-invasive system for identifying pigs’ instantaneous emotional states. This system could thus be used on-farm, and serve as a tool to professionals for controlling the welfare threats of pigs (e.g. piglets’ crushing, fighting or hunger). They could use it to monitor and improve pig’s welfare by minimizing stress and encouraging positive emotions.