Wrinkles and spots allow new diagnostic AI to estimate peoples’ ages
An artificial intelligence (AI) system developed by neuroinformatics engineers at the Institute for Neural Computation at Ruhr-Universität Bochum (RUB), Germany, can estimate a person’s age and ethnic origin almost as well as humans. The system is a “hierarchical neural network” that has learned to assess faces – particularly, the tell-tale signs of a normal aging process such as wrinkles, furrows and spots.
The researchers at RUB fed their system with several thousand photos of faces of different ages. The age was known in each case. But, the system ignored the features that varied from one picture to the next and took solely those features into consideration that change slowly.
“Think of it as a film compiled of thousands of photos of faces,” said Professor Laurenz Wiskott from the Institute for Neural Computation. “The system fades out all features that keep changing from one face to the next, such as eye colour, the size of the mouth, the length of the nose. Rather, it focuses on features that slowly change across all faces, like wrinkles.”
The system was then able to closely estimate the age of the people pictured in the photos and was only just under three and a half years off on average.
The system also reliably identifies ethnic origin: although the features characteristic of an ethnic group didn’t change quickly from image to image – they changed slowly, albeit by leaps and bounds – the new RUB algorithm nevertheless estimated the correct ethnic origin of the people in the photos with a probability of over 99%.
This means that the system outperforms even humans, who are the real experts in face recognition and interpretation, and could lead to promising facial analysis technology.
Category: Features, Technology & Devices
