Scientists invent “digital mask” to shield personal information

by Thanisha Kapur (’25) | October 7, 2022

Art by Gezan D’Souza (’23)

In the wake of the pandemic, people have become accustomed to wearing masks to protect themselves from the virus or to conceal their emotional state. Recently, researchers have developed a medical technology called the “digital mask” to anonymize the facial data of patients by censoring identifiable features while retaining qualities used for diagnosis.

Facial images can be useful for identifying signs of disease. For example, deep wrinkles around the eyes and forehead are associated with coronary heart disease, while abnormal eye movement can indicate poor visual function and cognitive developmental problems. However, facial images inevitably record other biometric information about a patient, including their race, sex, age, and mood. With the increasing digitization of medical records comes the risk of data breaches, which jeopardize patients’ sensitive personal information. 

In comparison to other patient information, it is especially difficult to make facial data anonymous without losing essential data. Common methods such as blurring and cropping identifiable areas may erase important disease-relevant information and cannot fully evade facial recognition softwares. Due to these privacy concerns, people often hesitate to share their medical data for public research or electronic health records, hindering the advancement of digital medical care. 

Professor Haotian Lin from Sun Yat-sen University explained, “During the COVID-19 pandemic, we had to turn to consultations over the phone or by video link rather than in person…. Patients want to know that their potentially sensitive information is secure and that their privacy is protected.”

Professor Lin and his colleagues were part of a team of scientists who helped develop a “digital mask,” which allows for an original video of a patient’s face to be inputted and outputs an altered video while discarding as much personal biometric information as possible. The “digital mask” does so through the use of a deep learning algorithm and 3D reconstruction. Deep learning extracts a patient’s facial features, while 3D reconstruction automatically digitizes the shapes and movement of the face based on these features. Converting the digital mask videos back to the original videos is extremely difficult because most of the necessary information is not retained.

The researchers then tested how effective the masks were in a clinical setting and found that diagnoses using the new technology were consistent with those based on the original videos. These results suggested that the reconstruction was precise enough for real-world medical use, and the risk of being identified was significantly lower in digitally-masked patients. 

The team also surveyed randomly selected patients within clinics to test their attitudes toward digital masks. Over 80% of patients believed the mask would alleviate their privacy concerns and expressed an increased willingness to share personal information if such a measure were implemented. 

Finally, the team confirmed that the digital masks could also evade potentially malicious Artificial Intelligence-powered facial recognition algorithms. Professor Patrick Yu-Wai-Man from the University of Cambridge described the implications of this finding: “Digital masking offers a pragmatic approach to safeguarding patient privacy while still allowing the information to be useful to clinicians… This could make telemedicine—phone and video consultations—much more feasible, making healthcare delivery more efficient.” With the invention of the digital mask and society’s rapidly advancing technology, researchers hope to soon overcome the remaining barriers and concerns related to privacy protection.

Categories: Science

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s