Research finds women have more expressive smiles than men
A University of Bradford study found that the dynamic movements of a smile differs between men and women. Women have more broad, expressive smiles and expand their mouths and lips more often than men. In recent years, researchers have used artificial intelligence to help identify information that faces can reveal such as age, ethnicity and health. For example, iPhone X’s recent launch of Animoji uses emojis to replicate facial movements and expressions. Similarly, at the University of Bradford, Professor Hassan Ugail led a study on gender classification based on the dynamics of a smile.
The study titled, “Is gender encoded in the smile? A computational framework for the analysis of the smile driven dynamic face for gender recognition,” published in Springer Link, used artificial intelligence to recognize gender. Similar research that used this technique includes the 2004 studies “Facial appearance, gender, and emotion expression” and “Smile line and periodontium visibility. Periodontal Pract.”
The difference, however, is that the previous research focused on gender classification with the use of artificial intelligence on single static images. The still images used in previous research were all different due to the lighting or pose of the individual. Ugail used the dynamics, or the underlying muscle movements used with every smile, of a smile to identify whether the smile belonged to a male or female.
In the study’s introduction, smiles were referred as “rich, complex and sophisticated facial expressions, formed through the synergistic action of emotions.” Researchers revealed in the introduction that they chose to analyze smiles because a smile can reveal various emotions. Instead of using appearance-based image analysis methods, researchers used an algorithm to detect gender.
Forty-nine different areas on the face were used in analyzing the dynamics of the face, and researchers focused on four distinct components of the face: spatial features, the dynamic triangular area of the mouth, geometric flow around the key parts of the mouth and the set of intrinsic features based on dynamic geometry of the face. In simpler terms, the algorithm tested the flow of the smile based on how much of the smile was formed, how far the smile stretched on the face and how fast it was formed.
To conduct this study, 109 participants’ smiles were recorded and analyzed. Compared to a 60 percent accuracy rate in still image facial recognition, the study using the dynamics of a smile resulted in 86 percent accuracy.
Additionally, this study coincides with previous observations that women tend to smile more expressively and wider than men. A 1995 study titled “Gender-based expectancies and observer judgments of smiling” and a 2016 study titled “Gender and emotion: theory, findings, and content” confirmed that women have more expressive smiles.
Emily Tung, a junior majoring in finance at Baruch, responded to this study by saying, “One would think since images stay put, they’d have more consistency in determining things like this.”
To add to her astonishment of this study, Tung said, “I never knew there was such a difference in the way males and females smile that it could determine gender.”
In retrospect, Tung also questioned, “Who knows what else our faces will play for technology in the future? I can only imagine. I would think it’d be most pertinent in the criminal field.”
On the other hand, Ivan Proshchenko, a junior majoring in economics at Baruch said, “Women in general smile differently than men, so this study does not surprise me. Women are more open, and more emotionally free than men. Men smile in more controlled and reserved smiles, and they don’t give smiles right and left.”
Overall, the study was advantageous since biometric identification tools always have room for improvement and more accuracy in the results of facial recognition tools. Additionally, given the previous studies on identifying gender based on still images, it is helpful to know that a dynamic approach provides results more closely matching the reality of the gender of the individual. Further studies conducted on this topic may want to include the accuracy of people who have undergone plastic surgery or identify as transgender.