@Misc{fdr_mods_00000316,
  author = 	{Wahmhoff, Johann
		and Traulsen, Imke
		and van Asten, Astrid
		and Krieter, Joachim
		and Dirksen, Neele
		and Diers, Sophie
		and Wutke, Martin},
  title = 	{SowPostureDS},
  year = 	{2026},
  month = 	{May},
  day = 	{27},
  keywords = 	{Computervision; Object Detection; YOLO; Sow husbandry},
  abstract = 	{SowPostureDS is a dataset designed to support the development of robust computer vision and artificial intelligence models for sow posture detection in farrowing environments. The dataset comprises 14.400 annotated images collected across three distinct housing environments, with balanced representation of the four posture classes lying lateral, lying, sitting and standing. The data includes both daytime and nighttime recordings. All images have been standardized by resizing to match a 1280 x 800 pixel resolution. The images are contained in .jpg format and the corresponding annotation informations are saved as .txt files. The dataset is provided in a .zip archive and contains the two subfolders' images and labels, which follow the standard directory structure commonly used for training YOLO models, enabling seamless integration into existing training workflows. The images and labels from the different environments are marked with corresponding prefixes (``ziss'', ``digi'' and ``inno''). For further information see the following publication. If you use this dataset, please reference: Wahmhoff, J., Traulsen, I., van Asten, A. L., Krieter, J., Dirksen, N., Diers, S., Wutke, M. (2026), SowPostureDS: A Multi-Class Image Dataset for YOLO-Based Detection of Sow Postures in diverse Farrowing Systems. Scientific Data},
  doi = 	{10.57892/100-316},
  url = 	{https://opendata.uni-kiel.de/receive/fdr_mods_00000316},
  url = 	{https://doi.org/10.57892/100-316},
  file = 	{:https://opendata.uni-kiel.de/servlets/MCRFileNodeServlet/fdr_derivate_00000320/SowPostureDs.zip:TYPE},
  language = 	{en}
}