What’s the news? Researchers have developed a technology called the digital mask (DM) to anonymise facial data of patients by erasing “identifiable features while retaining disease-relevant features relevant for diagnosis,” a paper published on Nature.com says. Further, the authors claim that the resultant anonymised data has the ability to even “evade recognition systems using artificial intelligence-powered re-identification algorithms.” Why it matters: India’s mission to digitise healthcare records and the rising adoption of telemedicine involves large scale collection of data. Storing facial images in medical records poses privacy risks due to the sensitive nature of personal biometric information that can be extracted from such images, the paper says. Privacy protection is important especially in context of data hacking and data leaking incidents. In the first half of 2022 alone, over 20 million health records were breached in the US. Moreover, India is on its way to draft guidelines on anonymisation of data and this paper's digital mask technology may be of interest to policymakers. How the technology works: The digital mask inputs an original video of the patient's face and outputs a de-identified three dimensional (3D) reconstructed video based on complimentary use of deep learning and 3D reconstruction. Deep learning helps in extracting parts of facial features 3D reconstruction automatically digitises the shapes and motions of 3D faces, eyelids and eyeballs based on the extracted facial features Have a look at the image below to get a better understanding of how the anonymised facial representation looks like: Courtesy: nature.com To develop…
