Clinicians using smartphones to capture images of patients’ eyes risk misdiagnosis if they base their decisions on objective data extracted from non-calibrated cameras, according to new research published by Anglia Ruskin University’s Vision and Eye Research Institute.
The study, published in the Nature journal Scientific Reports, looked at 192 images of eyes taken with three different smartphone cameras, two lighting levels and zoom levels of x10 and x6. The images were duplicated and one set was white balanced and colour corrected (calibrated) and the other left unaltered.
The photographs in were captured using autofocus mode with an Apple iPhone 6s, the Google Nexus 6p and the Bq Aquaris U Lite. The iPhone’s results were found to be significantly different from the other two devices when computing relative redness of each eye, and when compared to a clinician’s diagnosis. Once the images were calibrated, the differences between lighting levels and camera types on the smartphones were reduced by approximately 30%.
Lead author Carles Otero, of Anglia Ruskin’s Vision and Eye Research Institute said: ‘The use of smartphones in conjunction with ophthalmic equipment is becoming more and more widespread. Using smartphones is convenient and portable, meaning there’s no need to carry bulky equipment between sites.
‘However, this is the first time that the performance of three different smartphone cameras were evaluated in the context of a clinical application. Camera manufacturers have their own autofocus algorithms and hardware specifications, and this means different cameras can produce different results for the same scene. It is important that clinicians bear this in mind.
‘Our results show that while the clinician’s subjective evaluation was not affected by different cameras, lighting conditions or optical magnifications, calibration of a smartphone’s camera is essential when extracting objective data from images. This can affect both telemedicine and artificial intelligence applications.’