The 12th IEEE International Conference on Intelligent Systems (IS'24), held in Varna, Bulgaria, has awarded the prestigious Best Paper accolade to Sam Hennessey for his innovative research titled “Hierarchical Vs Centroid-Based Constraint Clustering for Animal Video Data.” Co-authored by Sam Hennessey, Francis Williams, and Lucy Kuncheva, the paper introduces an advanced comparison between hierarchical and centroid-based constraint clustering techniques for analysing video data of animal behaviour. Sam Hennessey is a PhD student at the School of Computer Science and Engineering, funded by the Artificial Intelligence, Machine Learning and Advanced Computing (AIMLAC) doctoral training centre, which is funded by UK Research and Innovation. Hennessey's work was recognised for advancing traditional data analysis in wildlife monitoring, with the potential to enhance conservation and behaviour prediction efforts.
The IEEE International Conference on Intelligent Systems (IS) is a prestigious annual conference organised by the IEEE Computational Intelligence Society, bringing together leading researchers, practitioners, and academics to discuss advances in intelligent systems and artificial intelligence.
The core of Hennessey's research lies in exploring two clustering methods — hierarchical clustering and centroid-based clustering on challenging animal movement data in video footage. Hierarchical clustering creates a nested grouping structure of the data. Centroid-based clustering, on the other hand, works by dividing data into clusters based on a calculated “centroid” or centre point for each group. None of the off-the-shelf approaches was able to recognise the individual animals within the length of the video footage. Hennessey’s team proposed a constrained clustering version of the two approaches over a sliding window along the video.
Hennessey's award-winning research has potential implications for wildlife and ecological studies, as clustering technologies like these can automate large-scale video analysis, making it feasible to monitor vast ecosystems with minimal human intervention. By making it easier to detect changes in animal behaviour patterns, these clustering techniques can help researchers identify early signs of environmental stress, behavioural shifts due to climate change, and other critical insights into animal welfare and ecosystem health.
“Receiving this award is a tremendous honour,” said Sam. “It reflects the hard work and collaboration that went into this research, and I’m grateful to the IEEE community for their support.”