Dental News - Smile research could lead to redesigning of facial recognition technology

Search Dental Tribune

Smile research could lead to redesigning of facial recognition technology

In a recent UK study highlighting the considerable difference in the way men and women smile, researchers mapped out 49 landmarks on the face, and then used the data to create an algorithm for facial recognition. (Photograph: University of Bradford)

Thu. 26. April 2018

save

BRADFORD, UK: According to a recent study, it is not only the obvious physical features that help to distinguish men from women, but also some subtler differences that are hiding in plain sight. Researchers from the University of Bradford have found that men’s and women’s smiles differ significantly, so much so that artificial intelligence (AI) is able to accurately assign sex based on no other information.

Existing methods of gender recognition within AI use a combination of geometric and appearance features which are both static and dynamic. This new research is the first of its kind to only use the dynamic movement of the smile to distinguish between men and women. Due to this, researchers believe it could become a next-generation biometric identifier, as it is not dependent on one feature, but on a set of intrinsic smile parameters unique to an individual, which would be very difficult to mimic or alter.

Lead researcher Prof. Hassan Ugail, Director of Centre for Visual Computing at the university, said: “Anecdotally, women are thought to be more expressive in how they smile, and our research has borne this out. Women definitely have broader smiles, expanding their mouth and lip area far more than men.”

Mapping out 49 landmarks on the face, with a focus on areas around the eyes, mouth and down the nose, the researchers were able to assess the physical changes as a person smiles through the movement of the muscles. In comparing smiles of both sexes in order to see if there was any significant difference, they were able to show that, out of the 109 participants (69 women and 40 men), the female smile was far more expansive than the male.

After gathering the data, the team created an algorithm using their analysis and tested it on video footage of the participants as they smiled. According to the results, the computer was correct 86 per cent of the time; however, the researchers believe this could easily be improved.

“Because this system measures the underlying muscle movement of the face during a smile, we believe these dynamics will remain the same even if external physical features change, following surgery for example,” said Ugail.

The study, titled “Is gender encoded in the smile? A computational framework for the analysis of the smile driven dynamic face for gender recognition”, was published in the Visual Computer on 5 March 2018.

Tags:
To post a reply please login or register
advertisement
advertisement