Longitudinal computer vision-based behavioural phenotyping of group-housed dogs for translational and welfare research

Continuous behavioural monitoring is essential for assessing animal welfare and individual behavioural variation, yet traditional observation-based methods are labour-intensive and prone to bias. Here, we developed a modular framework with computer vision for automated behavioural profiling of group-housed Beagles, integrating You Only Look Once (YOLO)-based models for pose estimation, individual identification, and enrichment detection. A ceiling-mounted infrared camera system enabled uninterrupted monitoring under standard kennel conditions. The identity-independent YOLO-POSE model achieved 99% precision, 97% recall, and an mAP50 of 0.98, while the YOLO classification model reached 95% accuracy for individual identification without external markers. Combining posture, identity, and environmental information enabled automated quantification of spatial use, locomotor activity, social interactions, corridor-oriented behaviour, drinking behaviour, and enrichment use. Overall, this framework captured distinct behavioural profiles of individual dogs from the same litter housed in the same kennel, which can be used to assess animal welfare and to evaluate environmental enrichments objectively. Together, we demonstrate a scalable and non-invasive approach for continuous individual behavioural profiling of group-housed dogs for computational ethology and translational research. By separating pose estimation from identity assignment, the framework enables flexible deployment across changing social groups and facilitates application to previously unseen individuals with minimal training.

Authors

Institutions

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-20
DOI
https://doi.org/10.64898/2026.11.19.767794
Primary Topic
Animal Behavior and Welfare Studies
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Longitudinal computer vision-based behavioural phenotyping of group-housed dogs for translational and welfare research

Henning Richter, Tom Lorimer, Paula Vázquez‐Pianzola, Miho Sato et al.
4 links
bioRxiv (Cold Spring Harbor Laboratory)
Animal Behavior and Welfare Studies
preprint

Longitudinal computer vision-based behavioural phenotyping of group-housed dogs for translational and welfare research

Henning Richter, Tom Lorimer, Paula Vázquez‐Pianzola, Miho Sato, Oliver Sturman, Andrea Campa, Kazhal Almasi, Preami Uthayavathanan, Lorena Item, Juval Gutknecht, Dominik N'Gouan
preprint en

Abstract

Continuous behavioural monitoring is essential for assessing animal welfare and individual behavioural variation, yet traditional observation-based methods are labour-intensive and prone to bias. Here, we developed a modular framework with computer vision for automated behavioural profiling of group-housed Beagles, integrating You Only Look Once (YOLO)-based models for pose estimation, individual identification, and enrichment detection. A ceiling-mounted infrared camera system enabled uninterrupted monitoring under standard kennel conditions. The identity-independent YOLO-POSE model achieved 99% precision, 97% recall, and an mAP50 of 0.98, while the YOLO classification model reached 95% accuracy for individual identification without external markers. Combining posture, identity, and environmental information enabled automated quantification of spatial use, locomotor activity, social interactions, corridor-oriented behaviour, drinking behaviour, and enrichment use. Overall, this framework captured distinct behavioural profiles of individual dogs from the same litter housed in the same kennel, which can be used to assess animal welfare and to evaluate environmental enrichments objectively. Together, we demonstrate a scalable and non-invasive approach for continuous individual behavioural profiling of group-housed dogs for computational ethology and translational research. By separating pose estimation from identity assignment, the framework enables flexible deployment across changing social groups and facilitates application to previously unseen individuals with minimal training.

bioRxiv (Cold Spring Harbor Laboratory)
University of Zurich (CH), ETH Zurich (CH)
Universität Zürich
Animal Behavior and Welfare Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon