Human-Aligned Super-Resolution for Facial Identification: Behavioural Evaluation, Bias Analysis, and Explainable AI

The Centre for National Training and Research Excellence in Understanding Behaviour (Centre-UB) at the University of Birmingham is offering a fully funded PhD studentship in partnership with VisionMetric Ltd, a leading provider of facial identification software for law enforcement. This interdisciplinary project, commencing October 2026, addresses the critical challenge of poor-quality CCTV footage in policing, where over 80% of real-world images are insufficient for reliable person identification. Generative AI-based super-resolution (SR) technologies, such as VisionMetric’s iREVEAL, have the potential to revolutionize facial identification by enhancing low-quality images, but their real-world impact on human and machine accuracy, as well as demographic bias, remains underexplored. The PhD will investigate how SR technologies affect both human and machine-based facial identification, combining behavioural experiments, machine learning, and explainable AI methods. Key research questions include: whether SR techniques improve human face identification accuracy, how SR-enhanced images influence machine recognition and the divergence between human and machine decisions, the equity of SR methods across demographic groups, and the potential for improving SR models using human perceptual insights. Students will receive comprehensive interdisciplinary training in behavioural experimental design, statistical modelling, computer vision, AI techniques, explainable AI, and responsible innovation. The project includes two placements at VisionMetric, providing hands-on experience with AI development pipelines and product development. This opportunity is ideal for candidates seeking to develop expertise at the intersection of psychology, AI, fairness, and forensic technology, preparing them for careers in academia, applied behavioural science, AI research, technology, or policy. The research addresses both societal risks and benefits of AI in high-stakes environments. Applicants should hold a 1st class or 2:1 undergraduate degree in Psychology, Cognitive Science, Computer Science, Neuroscience, Data Science, or a related field. An MSc in a relevant area is desirable but not essential. Experience in coding (Python, R, Matlab), behavioural experimentation, statistics, or machine learning is advantageous, with full training provided. Interest in human perception, AI ethics, or forensic science is especially encouraged. Centre-UB studentships cover tuition fees, a maintenance stipend, support for research training, and research activity support grants. Up to 30% of international applicants can be recruited each year due to UKRI funding stipulations. For further details and application instructions, visit Centre-UB Application Process and Call for Applicants . The application deadline is February 17, 2026, with interviews expected on March 16, 2026. Informal enquiries can be directed to Dr Melissa Colloff at [email protected] .