A new AI model that recognizes facial expressions.

Faces can tell a lot about our emotions and our mental state. Humans communicate nonverbally through facial expressions. We can often tell by looking at someone’s face what they are thinking, even if we don’t know how. Reading facial expressions can be very important in many situations. A teacher may use this technique to determine if students are interested or bored. Or a nurse might check to see if the patient’s health has worsened or improved.

Computers can recognize faces pretty well thanks to technological advances. It’s a different story when it comes to recognizing facial expressions. Researchers in the field artificial intelligence (AI), including convolutional neural network (CNN) are among those who have attempted to solve this problem. Face expression recognition requires complex neural networks that require extensive training and computational power.

To address these issues, Dr. Jia TIAN from Jilin Engineering Normal University (China) has developed a new CNN facial expression model. In an article in the Journal of Electronic Imaging the team was focused on finding a balance between training speed, memory use, and recognition accuracy of model.

Source:
https://techxplore.com/news/2022-07-written-ai-facial.html