The project, SEE-SAFE: Understanding and Reducing Visual Distortion in Clinical Image Transfer and Video Consultations, is among a range of newly funded studies addressing health and social care priorities across Wales.
Across health services, it’s increasingly normal for people to text or upload photos to their medical teams. But as Sam Altman of Open AI recently said in an interview, “Even like a photo you take out of your iPhone today, it's like mostly real, but it's a little not.” (interview with Cleo Abram, published 7 August 2025, titled Sam Altman Shows Me GPT 5... And What's Next.)
Despite this, decisions about diagnosis and treatment are being made based on photos taken by patient’s smart phones and shared. A whole range of factors from the lighting, the camera, AI processing, compression and the quality of the clinicians’ screen affect how accurate these images are, but these have not been formally investigated to date, nor are there guidelines for health services on which photos are and aren’t good enough for clinical decision making.
Bangor’s research will look at this area in three main ways, firstly by taking photos of patients with conditions such as leg ulcers or jaundice and seeing how much the images change according to the camera and lighting use, and if clinicians can still recognise clinical findings in these photos.
Secondly, by working with patient safety experts in Cardiff University to understand where misinterpretation of patient’s pictures might have led to harm.
Finally, by working with staff and patients in the health service to create guidance on how to get high-quality images, and on when images aren’t safe for decision making. Researchers on the project also hope to collaborate with the tech industry to create “medical photography” settings on phones which strip away the photos and highlight when photos aren’t good enough to use for your health.
Professor James Wang who brings his expertise in Imaging and Laser Micromachining to the project said “I'm excited to bring computer science and AI expertise into an area with such immediate implications for patient safety. Modern smartphones do far more than simply capture an image: software and AI can alter colour, contrast, sharpness and detail in ways that are largely invisible to the user. Our aim is to understand when those changes matter clinically, and then work with clinicians, patients and the technology industry to make medical images more reliable and safer to use.”
Dr Rebecca Payne, Director of the North Wales Centre for Primary Care Research said “Photos sent by patients now influence everyday decisions about diagnosis and treatment, but we still know surprisingly little about how faithfully those images reflect what is actually in front of the camera. A tiny change in colour or detail may not matter in a holiday snap, but in healthcare it could change a clinical decision. I’m really excited to be part of this research because it tackles a very real patient safety problem and could lead to practical guidance that makes remote care safer for patients and clinicians”
This piece of work forms an important part of Bangor’s growing expertise in digital health, bringing together the highly regarded school of computer science and engineering with the new medical school to address real world problems that matter both here in Wales but internationally too.
James, Rebecca and colleagues have previously published on this area in the Lancet Primary Care Payne RE, Woywodt A, Wang Z, Kirk UB, Brady M, Joishy M, King K, Mahdi A, Fischer Pedersen A, Huibers L. How a lack of visual fidelity in patient-generated images risks compromising care. The Lancet Primary Care. 2026;2(4):100120
Bangor researchers secure funding to make medical images safer
Bangor University researchers have secured funding from Health and Care Research Wales to investigate a growing patient safety concern: how reliably smartphone photos and video consultations capture the clinical details needed for diagnosis and treatment.