Consistent multi-animal pose estimation in cattle using dynamic Kalman filter based tracking
- Author
- Maarten Perneel (UGent) , Ines Adriaens (UGent) , Ben Aernouts and Jan Verwaeren (UGent)
- Organization
- Project
- Abstract
- Over the past decade, studying animal behaviour with the help of computer vision has become more popular. Replacing human observers by computer vision lowers the cost of data collection and therefore allows to collect more extensive datasets. However, the majority of available computer vision algorithms to study animal behaviour is highly tailored towards a single research objective, limiting possibilities for data reuse. In this perspective, pose-estimation in combination with animal tracking offers opportunities to yield a higher level representation capturing both the spatial and temporal component of animal behaviour. Such a higher level representation allows to answer a wide variety of research questions simultaneously, without the need to develop repeatedly tailored computer vision algorithms. In this paper, we first cope with several weaknesses of current pose-estimation algorithms and thereafter introduce KeySORT (Keypoint Simple and Online Realtime Tracking). KeySORT deploys an adaptive Kalman filter to construct tracklets in a bounding-box free manner, significantly improving the temporal consistency of detected keypoints. In this paper, we focus on pose estimation in cattle, but our methodology can easily be generalised to any other animal species. Our algorithm is able to detect up to 80% of the ground truth keypoints with high accuracy, with only a limited drop in performance when daylight recordings are compared to nightvision recordings. Moreover, by using KeySORT to construct skeletons, the temporal consistency of generated keypoint coordinates was largely improved, offering opportunities with regard to automated behaviour monitoring of animals.
- Keywords
- Pose estimation, Pose tracking, Adaptive Kalman filter, Behaviour monitoring, DAIRY-CATTLE, BEHAVIOR, IDENTIFICATION, NETWORKS, COWS
Downloads
-
publisher version.pdf
- full text (Published version)
- |
- open access
- |
- |
- 3.93 MB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-01JYPB617RDWGQAJ3644BTPCD3
- MLA
- Perneel, Maarten, et al. “Consistent Multi-Animal Pose Estimation in Cattle Using Dynamic Kalman Filter Based Tracking.” SMART AGRICULTURAL TECHNOLOGY, vol. 11, 2025, doi:10.1016/j.atech.2025.101014.
- APA
- Perneel, M., Adriaens, I., Aernouts, B., & Verwaeren, J. (2025). Consistent multi-animal pose estimation in cattle using dynamic Kalman filter based tracking. SMART AGRICULTURAL TECHNOLOGY, 11. https://doi.org/10.1016/j.atech.2025.101014
- Chicago author-date
- Perneel, Maarten, Ines Adriaens, Ben Aernouts, and Jan Verwaeren. 2025. “Consistent Multi-Animal Pose Estimation in Cattle Using Dynamic Kalman Filter Based Tracking.” SMART AGRICULTURAL TECHNOLOGY 11. https://doi.org/10.1016/j.atech.2025.101014.
- Chicago author-date (all authors)
- Perneel, Maarten, Ines Adriaens, Ben Aernouts, and Jan Verwaeren. 2025. “Consistent Multi-Animal Pose Estimation in Cattle Using Dynamic Kalman Filter Based Tracking.” SMART AGRICULTURAL TECHNOLOGY 11. doi:10.1016/j.atech.2025.101014.
- Vancouver
- 1.Perneel M, Adriaens I, Aernouts B, Verwaeren J. Consistent multi-animal pose estimation in cattle using dynamic Kalman filter based tracking. SMART AGRICULTURAL TECHNOLOGY. 2025;11.
- IEEE
- [1]M. Perneel, I. Adriaens, B. Aernouts, and J. Verwaeren, “Consistent multi-animal pose estimation in cattle using dynamic Kalman filter based tracking,” SMART AGRICULTURAL TECHNOLOGY, vol. 11, 2025.
@article{01JYPB617RDWGQAJ3644BTPCD3,
abstract = {{Over the past decade, studying animal behaviour with the help of computer vision has become more popular. Replacing human observers by computer vision lowers the cost of data collection and therefore allows to collect more extensive datasets. However, the majority of available computer vision algorithms to study animal behaviour is highly tailored towards a single research objective, limiting possibilities for data reuse. In this perspective, pose-estimation in combination with animal tracking offers opportunities to yield a higher level representation capturing both the spatial and temporal component of animal behaviour. Such a higher level representation allows to answer a wide variety of research questions simultaneously, without the need to develop repeatedly tailored computer vision algorithms. In this paper, we first cope with several weaknesses of current pose-estimation algorithms and thereafter introduce KeySORT (Keypoint Simple and Online Realtime Tracking). KeySORT deploys an adaptive Kalman filter to construct tracklets in a bounding-box free manner, significantly improving the temporal consistency of detected keypoints. In this paper, we focus on pose estimation in cattle, but our methodology can easily be generalised to any other animal species. Our algorithm is able to detect up to 80% of the ground truth keypoints with high accuracy, with only a limited drop in performance when daylight recordings are compared to nightvision recordings. Moreover, by using KeySORT to construct skeletons, the temporal consistency of generated keypoint coordinates was largely improved, offering opportunities with regard to automated behaviour monitoring of animals.}},
articleno = {{101014}},
author = {{Perneel, Maarten and Adriaens, Ines and Aernouts, Ben and Verwaeren, Jan}},
issn = {{2772-3755}},
journal = {{SMART AGRICULTURAL TECHNOLOGY}},
keywords = {{Pose estimation,Pose tracking,Adaptive Kalman filter,Behaviour monitoring,DAIRY-CATTLE,BEHAVIOR,IDENTIFICATION,NETWORKS,COWS}},
language = {{eng}},
pages = {{19}},
title = {{Consistent multi-animal pose estimation in cattle using dynamic Kalman filter based tracking}},
url = {{http://doi.org/10.1016/j.atech.2025.101014}},
volume = {{11}},
year = {{2025}},
}
- Altmetric
- View in Altmetric