Face Off

On the face of it, an MRI brain scan might not look much like a recognisable person, but removing facial features (pictured, bottom, with full scans above) is an essential step in ensuring that privacy is maintained when people’s scans are used in research. Altering images inevitably risks influencing their reliability, so researchers investigated how ‘defacing’ impacted human and automated quality assessments of a set of scans in an openly-available database. Trained human raters ranked the images as being of different quality when defaced – on average slightly higher – especially on lower quality scans. The bias was most pronounced in the most experienced ranker’s judgements, suggesting preconceptions that have accumulated over time. But automated computer quality assessment was unaffected by defacing. The researchers suggest ensuring quality checks happen before images are defaced to avoid bias creeping in at the early stage of processing that might affect analysis and research outcomes later.

Written by Anthony Lewis

Anthony Lewis

Writer

Ant was the web editor in the early days of BPoD, and has since worked with the UK’s top science organisations as a writer and multimedia producer. He now splits his time between freelance work and managing multimedia at The Lancet.

Website

• Image from work by Céline Provins and colleagues
• Department of Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland
• Image originally published with a Creative Commons Attribution 4.0 International (CC BY 4.0)
• Published in PLOS Biology, April 2025