Behavior Analysis of Ducks in Confinement Housing Systems: Protocol for a Scoping Review

The increasing global demand for meat products is posing a welfare risk to poultry. Precision livestock farming offers ways to balance production and animal welfare. Machine learning and computer vision are some of the most important research areas of precision poultry farming. Most studies in this field, especially behavior analysis, focus on chickens, as they make up most of the world's poultry population. The goal of this study is to make computer vision research more accessible on ducks, which are the second most prevalent poultry species globally. We have prepared a protocol for a scoping review to collect all studies detailing behavior analysis methods for commercially reared ducks. Until now, we have selected a list of 60 relevant studies through three rounds of screening. We describe the methods we will use to extract data from full texts and how we will make the results available for future computer vision research on duck behavior.