Congratulations to Richard Guest from the School of Engineering and Digital Arts on the award of an £288,334 grant from EPSRC entitled Hummingbird: Human-machine integration for biometric authentication in collaboration with the School of Psychology at the University of Southampton.
The aim of the HUMMINGBIRD project is to deliver a human-inspired framework for biometric analysis that improves on the separate capabilities of machine-based systems or human operators. Biometric analysis may be achieved within the limits of human capability by a dedicated human operator manually checking a sample against a reference to determine a match. Humans cannot complete this sort of task at volume, and they show particular cognitive biases, based on expectation, prevalence, or prior beliefs that can lead to decision errors. Machine algorithms alleviate the problems associated with cognitive load, fatigue and cognitive bias and can offer a scalable solution. However, they often only operate well within restricted parameters that the human operator can overcome. This project will explore the combination human and machine processes in order to realise the best of both approaches. The result will be the development of a proof-of-concept multi-expert, multi-modality, biometric recognition framework which we call ‘HUMMINGBIRD’