The physical process through which men and women smile differs sufficiently for artificial intelligence system to spot the difference and thus determine the gender of an individual. According to researchers from the University of Bradford in the U.K., a trained platform can automatically assign gender purely based on a smile.
While there are several automated gender recognition systems available, a downside with current methods is that they rely upon the use of static images and fixed facial features in order to generate result. This means they are less effective against a moving target. Taking a different approach, and working with dynamic activity, the researchers have successfully devised an automatic systems to distinguish between men and women.
To achieve this, the technologists mapped 49 distinguishing features around the face. The focus was primarily with the eyes, mouth and down the length of the nose. The researchers then used these scans to assess how the face changes as a smile is formed, in relation to underlying muscle movements. When we smile changes occur in terms of the distances between the different points together with the ‘flow’ of the smile (that is how much, how far and how fast the different points on the face move as a smile develops). In terms of the main gender difference, a woman’s smile tends to be more expansive.
The researchers developed an algorithm based on the analysis and this was tested using video footage of 109 people, with each person smiling. The artificial intelligence successfully determined the gender in 86 percent of cases. The researchers are working in improving the accuracy still further.
According to lead researcher Professor Hassan Ugail: “This kind of facial recognition could become a next- generation biometric, as it’s not dependent on one feature, but on a dynamic that’s unique to an individual and would be very difficult to mimic or alter.”
The approach taken for the new study is discussed in the journal The Visual Computer, with the peer reviewed research paper titled “Is gender encoded in the smile? A computational framework for the analysis of the smile driven dynamic face for gender recognition.”