Behavioural social network analysis to automatically detect health and welfare compromises in pigs
Supervisors:
Prof Ilias Kyriazakis, School of Biological Sciences and Institute for Global Food Security (Queen's University Belfast)
Dr Niall McLaughlin, School of Electronics, Electrical Engineering and Computing Science (Queen’s University Belfast)
Prof Craig Michie, Department of Electronic and Electrical Engineering (University of Strathclyde)
Summary:
The aim of the studentship is to develop and apply the novel approach of social network analysis to automatically quantify changes in pig behaviour to detect health and welfare compromises. The hypothesis is that changes in group behaviour precede clinical signs of health disruptions and thus have diagnostic value that can be used to create an automated early warning system. We will utilise videos of situations where the behaviour of pigs has been disrupted either artificially or naturally, eg through an infection. We will build on existing video-based tracking methods developed by the collaborative team, using computer vision and deep learning, and extend these methods to allow for real-time, long-term operation. We will consider the application of graph theory algorithms to reason about social interaction networks given long-term trajectory information and detected events.
This is a cross disciplinary project. Students will be offered the chance to attend modules of either the MSc course in Animal Behaviour and Welfare (QUB) or MSc in Data Analytics (QUB). At the associated University (Strathclyde) the student will interact with a variety of practitioners of agri-technologies. The candidate will acquire a rare combination of technical, scientific, and hands-on skills that are highly valued in both industry and academia. The student will also benefit from association through his/her supervisors the Global Innovation Institute (GII, QUB) aims to develop transformational digital and data analytics tools to advance the wellbeing of animals.