ICT develops virtual-form psychologist

Researchers from USC’s Institute for Creative Technologies are slated to complete the development of their avatar system, which detects signs of depression, by September 2013.

“We [asked], ‘How can technology help clinicians in their work?’” said Louis-Philippe Morency, one of the principal researchers on the project.

Morency and his fellow researcher on the project, Albert Rizzo, answered this question with SimSensei, a system designed to detect signs of psychological distress based on analysis of facial movements and voice indicators.

Morency and Rizzo said they hope the prototype will see real-world use within the next two years, and they look forward to implementing this technology because it can be a valuable tool for sensitive patient cases.

Morency and Rizzo teamed up because they were willing to support one another through their research by sharing their expertise. Morency’s specialty lies in the technological side of the project, while Rizzo offers specialized skills in psychology.

Morency credits a large part of the project’s origin to Rizzo and his previous work in psychology, particularly with post-traumatic stress disorder.

“[Rizzo] had the idea of taking the technology and applying it to the medical world,” Morency said. “He’s been working really hard.”

Multiple scientists have contributed to this product over the past decade. They have pooled together their individual research, information and technology in order to create the avatar.

“The avatar itself had been started about 10 years ago because the technology that we were using now has been built over the last 10 years so, again, this is a team effort,” Morency said. “We started [SimSensei] a year and a half ago in October 2011. The ‘team-effort’ attitude is particularly evident in the breadth of expert fields used to work on SimSensei, with disciplines involving 3-D animation, writing and programming.”

Some USC students were involved in the project at a preliminary level. They conducted research to assist Morency, Rizzo and their colleagues but did not work directly on the technology or prototype. Morency said many postdoctoral researchers were particularly essential in the later stages of the project.

As the prototype release date comes closer, Morency said the team currently faces three major challenges before debuting their optimal product.

“One [challenge] is we developed a lot of the indicators of depression from the face and from the voice and integrating all of these, which is called mutual fusion,” Morency said. “The second big challenge is really about dialogue, which is the human deciding what to say next and giving feedback and listener feedback and these two aspects of dialogue. And [third], getting it right, which humans are extremely good at — first, the timing of things, and second, giving the right feedback and empathy at the right time.”

A pragmatist at heart, Morency also said that this first prototype might be imperfect, but its potential benefits can be very impactful for future patients who might use it.

“We’re realistic, and we’re not going to solve all of [the problems]. But we’re looking to push it forward,” Morency said. “We’re keeping it together.”