TY - CONF

T1 - Hierarchical Vs Centroid-Based Constraint Clustering for Animal Video Data

AU - Hennessey, Sam

AU - Williams, Frank

AU - Kuncheva, Ludmila

PY - 2024/8/29

Y1 - 2024/8/29

N2 - The identification of animals from video footage isan important ecological pursuit. It presents a labour-intensivetask, requiring experts to invest considerable time analysing videorecordings. While identifying humans from video is a mainstreamresearch quest, much less has been done for recognising animals’identities. This paper explores and contrasts the effectivenessof hierarchical and centroid-based constraint clustering methodsacross five manually annotated video datasets containing animals.We aim to determine the most suitable methodology for integrat-ing into a fully autonomous pipeline for animal re-identification.Our experimental findings indicate that, contrary to expectation,online hierarchical constraint clustering surpasses centroid-basedconstrained clustering.

AB - The identification of animals from video footage isan important ecological pursuit. It presents a labour-intensivetask, requiring experts to invest considerable time analysing videorecordings. While identifying humans from video is a mainstreamresearch quest, much less has been done for recognising animals’identities. This paper explores and contrasts the effectivenessof hierarchical and centroid-based constraint clustering methodsacross five manually annotated video datasets containing animals.We aim to determine the most suitable methodology for integrat-ing into a fully autonomous pipeline for animal re-identification.Our experimental findings indicate that, contrary to expectation,online hierarchical constraint clustering surpasses centroid-basedconstrained clustering.

KW - Constrained clustering

KW - Hierarchical Clustering

KW - Animal re-identification

KW - Pairwise constraints

KW - Semi-supervised learning

M3 - Paper

T2 - 2024 IEEE 12th International Conference on Intelligent Systems (IS)

Y2 - 29 August 2024 through 31 August 2024

ER -