Disney Research has developed a neural network, a computer system based on the human brain and nervous system, designed to track the facial expressions of audiences in dark theaters. Disney researchers recently tested the new method, known as factorized variational autoencoders or FVAEs, and found that it could reliably predict viewers reactions for the rest of the movie after only observing them for a few minutes — a feat that few parents can achieve with their kids. (Unless it’s right before Bambi’s mom dies.)
“The FVAEs were able to learn concepts such as smiling and laughing on their own,” Zhiwei Deng, a Ph.D. student at Simon Fraser University and lab associate at Disney Research, told Phys.org. “What’s more, they were able to show how these facial expressions correlated with humorous scenes.”
Disney’s findings, presented at IEEE Conference on Computer Vision and Pattern Recognition in Honolulu on July 22, include analysis of 150 showings of several Disney movies such as Zootopia and Star Wars: The Force Awakens, 3,179 people, and over 16 million expressions. Studios analyzing the reactions of test audiences is nothing new, but having a computer shoulder that responsibility outperformed traditional methods significantly, through a system of pattern recognition.
It’s a two-part process. First, the FVAEs look for audience members who exhibit similar expressions throughout the movie. Then the computer uses this data to calculate stereotypical reactions from entire audiences Through this, the computer learns a range of general facial expressions, like smiles, and can determine how audiences are more likely to react to a given movie based on correlations between audience members.
“It’s more data than a human is going to look through. That’s where computers come in–to summarize the data without losing important details,” Disney Research scientist Peter Carr explained to Phys.org.
Parents can rest assured that this technology is just being tested and no one is using it on your kid (yet). Still, this is one of many examples where Disney is gaining ground in the race to utilize AI technology for optimal results in media. The company’s long-standing relationship with the Computer Science department at Carnegie Mellon along with recent developments suggests that the studio isn’t slowing down. While that may be a lot to wrap your brain (or analog neural network) around, it could help Disney reshape movies to people’s liking in real time. If that means your kid doesn’t have to look into Johnny Depp’s shark eyes as much, it could be good for the whole family.