AI can identify the gender of men and women through smiles

A new study shows that artificial intelligence can accurately characterize male and female gender. This is because women's smiles are wider than men. This finding allows researchers to create an AI algorithm that can determine gender by up to 86%. Not only that, the study also offers interesting questions about the ability of AI to solve complex sexual and human sexual problems.

The head of the research, Hassan Ugail, said: "Women are supposed to be more expressive in the way they smile, and our research also demonstrates this. Women are sure. have a wider smile, widen his mouth and lips more than men. ".

This method is based on the movement of the face when smiling by tracking the movement of 49 points on the face. The "landmarks" are mostly close to the eyes, mouth and nose. The research team analyzed 109 people: 69 women and 40 men.

Professor Ugail explained: "Because this system evaluates the movement of the facial muscles in a smile, we believe these signs will remain the same, even when the external body characteristics changes, such as after surgery, this facial recognition can become the next generation biological identification technology, because it does not depend on any one feature. ".

However, this study also raises questions about gender identity and transgender people, the team hopes to study more about this issue. They also point out that gender awareness plays an important role in "face recognition, age, ethnicity, identity analysis, video surveillance and intelligent computer interaction with people".

Not only that, separate research has shown that current facial recognition technology tends to identify men if facial skin tends to be a little whiter.

But working together with similar technology, it can help make facial recognition software more sophisticated.

The team thinks that facial expressions can help identify gender, including "surprise, fear, anger and disgust".

This study is being carried out at the University of Bradford in England. It has been published in Visual Computer: International Journal of Computer Graphics.