Researchers have created a ‘digital mask’ that allows facial images to be stored in medical records while preventing biometric information from being extracted.
The scientists used 3D reconstruction and deep learning algorithms to erase identifiable features from facial images while retaining potentially disease-indicating features.
Professor Haotian Lin from Sun Yat-sen University said:
“During the COVID-19 pandemic, we had to turn to consultations over the phone or by video link rather than in person.
“Remote healthcare for eye diseases requires patients to share a large amount of digital facial information.
“Patients want to know that their potentially sensitive information is secure and that their privacy is protected.”
Wrinkles around the eyes are significantly associated with coronary heart disease, while abnormal changes in eye movement can indicate poor visual function.
However, facial images also record other biometric information about the patient, including their age, sex, race and mood.
As medical records are being increasingly digitised, patient data is at greater risk of being breached than ever before.
While most patient data can be anonymised, it is difficult to mask facial data while retaining essential information needed for diagnosis.
Common methods such as blurring may lose important disease-relevant information and do not always evade facial recognition systems.
To develop the mask, the researchers first used deep learning to extract features from different facial parts.
They then used 3D reconstruction to digitise the shape and movement of 3D faces, eyelids and eyeballs based on the extracted facial features.
The scientists then tested how useful the masks were in clinical practice by showing 12 ophthalmologists digitally-masked or cropped images and asking them to identify the original from five other images.
The ophthalmologists correctly identified the original from the digitally-masked image in 27 per cent of cases.
Meanwhile, for the cropped figure, they were able to do so in 91 per cent of cases.
However, this is likely to be an over-estimation as, in real situations, an individual would likely have to identify the original image from a much larger set.
The team also surveyed randomly-selected patients attending clinics to test their attitudes towards digital masks.
More than 80 per cent of respondents believed the digital mask would alleviate their privacy concerns and expressed an increased willingness to share their personal information if the mask was implemented.
Finally, the team confirmed that digital masks can also evade AI-powered facial recognition algorithms.
Professor Patrick Yu-Wai-Man from the University of Cambridge said:
“Digital masking offers a pragmatic approach to safeguarding patient privacy while still allowing the information to be useful to clinicians.
“At the moment, the only options available are crude, but our digital mask is a much more sophisticated tool for anonymising facial images.
“This could make telemedicine – phone and video consultations – much more feasible, making healthcare delivery more efficient. If telemedicine is to be widely adopted, then we need to overcome the barriers and concerns related to privacy protection.
“Our digital mask is an important step in this direction.”
