Release date: 2015-11-12
In 1872, Charles Darwin pointed out in his book "Emotional Expressions of Humans and Animals" that in all body parts, the face is most valued because it is the main part of making expressions and making sounds.
Scientists have found a lot of evidence that some facial expressions are innate, but there are many complex social and cultural factors to consider. A 2012 study showed that Western-specific facial expressions are associated with six main emotions (happiness, sadness, surprise, fear, disgust, and anger), while the Orientals prefer eye contact.
Today, machine learning offers scientists a new way to explain subtle changes in the face. A new system called MultiSense tracks people's head position, direction of the eyes, and the tendency of the body to move in real time. These details can provide us with a lot of information. For example, observing someone's nose and eyebrows can help us distinguish between a true smile and a disguised smile. “For most of the time, the expressions of others we saw were just a mask that was politely dressed,†said Louis Philip Morensi, assistant professor at the School of Computer Science at Carnegie Mellon University.
“Some people smile to respond to other people’s smiles, so smiles change because of emotional and social conditions.â€
Moranci and his colleagues are particularly interested in using machine learning to study the relationship between facial expression and mental state in patients with depression. At the moment, they have had unexpected results. People with depression and non-depressed people laugh as often, but they smile differently—the smiles of depression patients last longer.
The facial expressions of people with depression of different genders are also different. In a study at the University of Southern California, Morensi and three other researchers found that people with depression had more frowns than non-depressed people, but in the female population, the result was exactly the opposite: suffering from depression Women have fewer frowns than women who do not have depression.
“There is a bit more interesting,†Morensi said. “That is, the results of these studies are in line with social norms.†For example, people tend to want women to smile more. “Is this related to culture? Is it related to local and national differences? Or is it a global phenomenon? Or is there another factor that we have not yet discovered?â€
From a medical perspective, this technology means that machine learning can help human doctors track the long-term situation of patients. In short, Morensi believes that MultiSense can compete with professional clinicians.
However, there are other uses for this type of technology. For example, the US military has invested in this research. The Department of Defense wants to use this facial recognition platform to treat soldiers suffering from post-traumatic stress disorder (PTSD). In addition, it has a long-term goal - to understand and predict people's behavior through this technology. To this end, the Department of Defense began building a vast database of facial expressions decades ago.
Morensi is excited about the development of the study, but he also urges people to be vigilant, emphasizing that these tools are best suited to the treatment of patients with depression, rather than as a diagnostic tool. It turns out that these human expressions are subtle and difficult to quantify:
“I personally think that this technique is best used to help doctors diagnose, not just rely on it for diagnosis. We also need to consider many ethical issues.â€
Source: translation - èŒèšª five-line spectrum
Yancheng Rongtai Labware Co.,Ltd , https://www.shtestlab.com