Trio of facial recognition standards are updated
An article published in the journal Standards Engineering calls out developing ASTM International standards for use in facial recognition.
The article is not about newly proposed standards. Instead, it is a progress report on developing standards in an industry in critical need of them.
The paper, “Understanding the Role of Facial Recognition Standards,” is available free on LinkedIn.
One standard being worked on would aid facial comparison examiners. Its objective would be “to describe the relative physical stability of facial features of adults when assessing the observed component characteristic withing a single living subject.”
A key aspect of the standard would be to address factors affecting the stability of facial features in an adult. The features would be ranked by how stable they are — low, medium and high. Less-stable features would be discounted in image searches and biometric comparisons.
The second standard discussed would guide role-based training in facial comparison. This one is all about setting basic-through-advanced objective criteria that competent professionals would meet.
A third proposed standard would be called Image Processing to Improve Automated Facial Recognition Search Performance. In this case, the guidance would standardize how a facial examiner processes a probe image so that an investigative lead can be maximized.
While some probe images need no massaging to be most useful, many do. Those might need manual processing within a facial recognition system by a trained examiner. The goal would be to accentuate features without altering them in the massaging process.
The IEEE also began working on facial recognition standards to protect against potential social harms in 2019.