FB pixel

Slashing glances + head-swiveling = danger. Biker biometrics could help city planners cut injuries

Slashing glances + head-swiveling = danger. Biker biometrics could help city planners cut injuries

Claims that AI could solve most problems can be a bit rich, but one factoid about federal rules for improving safety at urban street crossings make algorithmic options sound very good.

An article in ScienceDaily about the use of biometric systems in safety roles notes that federal rules require some extreme evidence before making street improvements.

There needs to be 90 to 100 pedestrians crossing a given spot every hour before regulations kick in to add, say, a traffic light to protect the walkers.

Or a minimum of five deaths attributed to vehicles mowing down pedestrians in one year. Those are some hard facts.

So, when a 2018 study just published in Accident Analysis & Prevention claims that eye-tracking surveillance systems based on facial recognition could flag dangerous locations without a single killing, the idea appears worth exploring.

The idea is to perform cluster analysis of stress indicators in cyclists’ eyes using biometric systems as bikers ride through Philadelphia. It comes out of research by University of Pennsylvania, Swarthmore College and Carnegie Mellon University scientists.

Algorithms would be set to spotting “lateral head movements, gaze velocity, and off-mean gaze distance” to gauge perceived risk captured by Tobii eye-tracking glasses. The glasses would have eye-facing and street-facing cameras and a gyroscope to harvest head and eye movements 100 times a second.

Future researchers may want to look at how wide a cyclists eyes become, which could indicate the size of oncoming pot holes.

Related Posts

Article Topics

 |   |   |   |   |   | 

Latest Biometrics News


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

The ID16.9 Podcast

Most Read This Week

Featured Company

Biometrics Research

Biometrics White Papers

Biometrics Events

Explaining Biometrics