FACE AI
Research type
Research Study
Full title
FACE AI: Automated Detection of Oculofacial Parameters Using Artificial Intelligence
IRAS ID
319431
Contact name
Swan Kang
Contact email
Sponsor organisation
Moorfields Eye Hospital Foundation Trust
ISRCTN Number
ISRCTN17529873
Duration of Study in the UK
1 years, 7 months, 9 days
Research summary
Examining the face can inform doctors about the systemic health of patients such as stroke, facial nerve palsy, Parkinsonism, head and orbital cancers, and chronic diseases with multi-morbidity. Doctors assess the face by inspecting and measuring facial features with a ruler. Despite this being the traditional method, it may not always be entirely accurate and there can be differences in quantitative measurements obtained by different doctors.
Video consultations have been vital in continuing the delivery of safe patient care during the pandemic, however, its main limitation is the inability to obtain facial measurements over video.
Advances in artificial intelligence (AI) have allowed machines to make more sense of pictures and videos shown to them. Our team has shown that AI can be used to examine faces, and measure at millimetre accuracy fine details like the drooping of an eyelid.We intend to develop and validate an AI algorithm that can examine faces and eyelids over video consultation. To train the AI, we will use data from videos of Moorfields Eye Hospital patients and healthy individuals to develop a software that can be used as a plug-in to current video consultation platforms. Facial measurements obtained by AI will be compared with those obtained by doctors to ensure the AI is reliable.
AI-assisted technology will allow patients to be seen in the comfort of their own home, reducing the burden in the already overstretched NHS outpatient services and will empower patients to monitor their own conditions, where serious cases can be automatically prioritised while reassurance is given to those that are stable or improving.REC name
London - West London & GTAC Research Ethics Committee
REC reference
23/LO/0355
Date of REC Opinion
26 Apr 2023
REC opinion
Favourable Opinion