Automated monitoring of animal welfare has largely targeted negative indicators, leaving positive welfare behaviours such as play underexplored. To address this gap, we present PlayClass, a pipeline for play-behaviour classification in poultry from top-down pen video. The pipeline leverages long-duration tracking with SAM 3 via YOLO-guided chunk boundaries to minimise identity errors in point-based prompting, and frozen embeddings from image and video foundation models for play action classification. Although handcrafted motion features from tracked masks alone achieved competitive accuracy, V-JEPA 2.1 consistently outperformed all other backbones across model scales, reaching 77.0 macro-averaged F1 when combined with handcrafted features. Despite this result, the dataset remains challenging due to play sub-types sharing similar kinematic profiles with non-play and inter-bird occlusion. Overall, our work provides encouraging evidence towards automated frameworks for play behaviour classification in poultry.
Cite
@article{arxiv.2605.27304,
title = {PlayClass: Automated Play Behaviour Classification in Poultry},
author = {Prince Ravi Leow and Neil Scheidwasser and Rebecca Oscarsson and Per Jensen and Samir Bhatt and David Alejandro Duchêne},
journal= {arXiv preprint arXiv:2605.27304},
year = {2026}
}