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

University of Birmingham - School of Psychology

The Centre for National Training and Research Excellence in Understanding Behaviour (Centre-UB) is inviting applications for a Doctoral Studentship in association with our collaborative partner VisionMetric Ltd (Vision Metric - Fast, Accurate, Cost Effective) to start in October 2026.

CCTV footage is a predominant and crucial source of evidence in policing, with an estimated 21 million cameras operating in the UK. Yet more than 80% of real-world footage is too poor in quality to support reliable person identification. This severely limits investigative success and leaves offenders unidentified.

Generative AI–based super-resolution (SR) technologies—such as VisionMetric’s iREVEAL—promise transformative gains by enhancing low-quality facial images. However, there is little scientific evidence on whether these tools improve human accuracy, how they affect machine recognition, and whether they introduce demographic biases.

This interdisciplinary PhD will investigate how generative AI–based super-resolution (SR) technologies influence human and machine-based facial identification. The PhD will combine behavioural experiments, machine learning, and explainable-AI methods to answer questions:

1. Do SR techniques improve human face identification accuracy?
2. How do SR-enhanced images affect machine-based facial recognition, and where do human and machine decisions diverge?
3. Do SR methods perform equitably across demographic groups?
4. Can SR models be improved using human perceptual insights?

This project provides extensive interdisciplinary training from subject experts and industry, including in behavioural experimental design and statistical modelling; computer vision and AI techniques; explainable AI and human–machine comparison methods; and responsible innovation.

The student will work closely with VisionMetric, which is a leading SME supplying facial software to police forces in over 30 countries. Two placements at VisionMetric will provide hands-on experience with AI development pipelines and product development.

This is an exceptional opportunity to build a skillset spanning psychology, AI, fairness, and forensic technology, positioning the candidate for careers in academia, applied behavioural science, AI research, technology, or policy.

The project addresses both the societal risks and potential benefits of AI in high-stakes environments.

We are looking for a highly talented and dedicated student with a 1st class or 2:1 undergraduate degree in Psychology, Cognitive Science, Computer Science, Neuroscience, Data Science, or a related field. An MSc degree in a relevant area is desirable though not necessary. Experience in coding (e.g., Python/R/Matlab) and experience in behavioural experimentation, statistics, or machine learning is desirable but full training will be provided. Applicants with an interest in human perception, AI ethics, or forensic science are especially encouraged.

To be considered for this PhD, please follow the instructions here: https://www.centre-ub.org/studentships/application-process/

Application deadline: February 17 2026

Interviews for this studentship are expected to take place on March 16 2026.

Centre-UB studentships cover tuition fees, a maintenance stipend, support for research training, as well as research activity support grants. Due to funding stipulations set by UKRI, we are able to recruit up to 30% of international applicants to the cohort each year. You can find further details at https://www.centre-ub.org/studentships/call-for-applicants/

Informal enquiries about the project prior to application can be directed to Dr Melissa Colloff (m.colloff@bham.ac.uk).