Facial recognition is a non-invasive method for analysing facial characteristics and changes in facial expressions. There is evidence to suggest there may be a link between physiological health and facial expression patterns. Many medical conditions are associated with specific facial characteristics, with some related to inherited genetic conditions and others that have been acquired as a result of medical conditions such as stroke, nerve injuries and dementia. If successful, this technology could prove useful for studies evaluating the monitoring of acute or chronic illness. The proposed study will ask participants to do a maximal exercise test that will challenge the body, as a proof-of-principle, to see how this correlates with facial expressions. Depending on the study outcomes and future work, it could prove to be a useful tool and clinical application in healthcare for evaluating and tracking patient health and well-being. The investigators want to conduct a proof-of-principle study where physically stressed young and healthy participants (in the form of a maximal exercise test), whilst recording facial expressions, will be used to determine the relationship between physiological stress and changes in facial parameters.

For this pilot study, no specific sample size calculation is required for this pilot study as this will recruit 15 participants for proof of concept. The outcomes from this research will lead to the identification of measurement parameters to be evaluated in future studies concerning facial characteristics and their parameters. This will inform the next steps in the application of the process to further healthy participants with the ultimate aim to translate this to research on a range of patient populations in elective and acute medical settings.

Video and still image capture will be undertaken using a camera mounted on a tripod in front of the participant whilst they are exercising. The data analysis mainly consists of four processing stages, including facial landmark detection, facial region identification, region-based facial analysis and data correlations, as illustrated in the diagram below. The facial landmark detection can label a total of 68 points in each face, which are in the predefined locations. Based on the facial landmark, the region identification can segment the face into different areas, such as brows, eyes, chin, etc. Then, the region-based analysis can be applied to facial data and extracts the geometry, colour, emotion and temperature information of the individual facial region and calculates their variations against time. Finally, vital sign reading and region-based facial data are analysed together by using statistical tools in order to identify their correlations and demonstrate how facial appearances change in different scenarios.

Lancaster University is the data owner/controller, and UCLAN will follow Lancaster's requirements of data management when the data is processed. The data will be transferred on an encrypted hard drive and stored on secure online storage by UCLAN data protection rules. The data will only be accessible to the FACIAL study collaborators and statisticians. Whilst sociodemographic and physiological data can be anonymised, the video recordings of facial expressions cannot. The two sets of data will be provided with separate research codes so that protected data cannot be linked to identities by anyone other than the study team. Processing of protected data such as ethnicity and biometric data is central to the research question and ultimately, might have efficacy in the treatment of high dependency patients in hospital care.

The FACIAL Study protocol has been formulated and reviewed by the FACIAL Steering Committee, and research will be carried out in line with the University of Central Lancashire and Lancaster University code of ethics. The FACIAL study will be carried out by the FACIAL steering committee as a collaboration between East Lancashire Hospitals NHS Trust, Lancaster University and the University of Central Lancashire.

The findings from this study will be presented at national or international conferences. Any facial images used for publication will be obtained with additional consent from the individuals. Manuscript(s) will be prepared for submission to a peer-reviewed journal.

Inclusion Criteria:

• Aged 18 to 35
• Male or female
• Capable of completing a maximal exercise test
• Answer "NO" to all variables in the Physical Activity Readiness questionnaire (PAR-Q)

Exclusion Criteria:

• Musculoskeletal injury
• Cardiovascular disease
• Neurological disease, stroke, or other disease that may affect facial expressions

Healthy participants from the general public, inlcuding students and staff from the university.

This pilot study is a proof-of-principle study where we physically stress young, healthy participants (in the form of a maximal exercise test), whilst recording facial expressions, to determine the relationship between physiological stress and changes in facial parameters. The code developed from this pilot will be used as part of a PhD project at UCLan and is therefore novel. Individual participant data will not be shared externally at this time.