Help children with autism to identify facial expressions of other people. That is the objective of an initiative that the Stanford University School of Medicine and Google have developed together with their augmented reality glasses, Google Glass, to which they introduced an application that used artificial intelligence (AI), learning automatic or machine learning (ML) and facial recognition.
For the study, published in the journal Digital Medicine, an application designed by the University was used, which provides real-time information about the facial expressions of other people to a child using Google Glass. As the autistic child interacts with others, the application identifies and names their emotions through the speaker or the Google Glass screen. After one to three months of regular use, parents of children with autism who used the device reported that their children made more eye contact and interacted better with others.
The researchers named the new therapy “Superpower Glass,” and it is based on behavior analysis, a well-studied autism treatment in which a clinician teaches emotion recognition using structured exercises, such as picture cards depicting faces with different emotions. .
The Stanford University School of Medicine team that led the study built a smartphone app that used machine learning to recognize eight core facial expressions: happiness, sadness, anger, disgust, surprise, fear, neutrality, and contempt. The app was trained with hundreds of thousands of photos of faces showing all eight expressions, and it also had a mechanism to allow the people involved in the study to calibrate it to their own “neutral” faces if necessary.
Fourteen families, each with an autistic child between the ages of 3 and 17, tested the Superpower Glass at home for 10 weeks. And they used the therapy for at least three 20-minute sessions per week with a facial recognition program that could be used in “free play” mode, in which children used Google Glass while interacting or playing games with their family. The software gave the autistic child a visual or auditory cue each time it recognized an emotion on someone’s face in the field of vision.
The Superpower Glass also allowed two other game modes. In one of them the father had to put a facial expression of those mentioned above, and the child had to identify it. In the third game mode, called “catch the smile”, the children gave another person clues about the emotion they wanted to provoke.
Twelve of the families who participated in the initiative reported that their autistic children made more eye contact after receiving treatment. A few weeks after the test, some of the children began to realize that people’s faces have clues to their feelings.
