Biometrics utilises the anatomy, physiology and behaviour of living beings for the purpose of identification, authentication or related analyses. Computer vision offers a non-intrusive way to build biometric systems. Here in Bristol we are particularly focussing on the application of visual biometrics to animals, contributing to an interdisciplinary field which is rapidly evolving now.
Animal Biometrics as a domain was defined by us and our collaborators in the TREE2013 paper as: ‘Animal biometrics is an emerging field that develops quantified approaches for representing and detecting the phenotypic appearance of species, individuals, behaviors, and morphological traits. It operates at the intersection between pattern recognition, ecology, and information sciences, producing computerized systems for phenotypic measurement and interpretation. Animal biometrics can benefit a wide range of disciplines, including biogeography, population ecology, and behavioral research. (…) However, to advance animal biometrics will require integration of methodologies among the scientific disciplines involved. Such efforts will be worthwhile because the great potential of this approach rests with the formal abstraction of phenomics, to create tractable interfaces between different organizational levels of life.’
Many applications in conservation, animal welfare, smart farming, and the behavioural sciences rely on the localisation of species members, the identification of individuals, and the recognition of specific activities and behaviours. Thus, part of our research is focussed on these aspects which we applied to a variety of target species.
Facial recognition in humans is well studied, however, facial biometric applications in animals are rare. We work with our collaborators to advance this form of identification in primates in particular. Potential use cases range from visitor engagement in zoo exhibits to manual photography and camera trap analysis for conservation in the wild.
Despite the fact that computer vision has advanced significantly over the last decade, recognition of animals and natural structures in natural environments still poses a challenge to systems due to visual complexity, camouflage, clutter, and occlusion. We are working on techniques to improve the robustness of recognition techniques ranging from active aerial vision solutions to deep learning detectors defying heavy occlusion and noise.