Emotient Granted Patent to Automate Facial Expression Analytics

New Patent Covers Crowdsourcing of Training Data for Deep Learning System

May 21, 2015, 08:05 ET from Emotient

SAN DIEGO, May 21, 2015 /PRNewswire/ -- Emotient, whose emotion analytics services help organizations better understand and engage with customers, today announced a newly issued patent on automated machine learning of facial expressions.  

The field of emotion analytics is gaining recognition as a much more precise way to measure individual and crowd reaction to an experience such as an advertisement, a live event or a service interaction. Current methods are flawed: surveys require that respondents accurately remember and articulate past emotional states; focus groups require subjective interpretation of feedback and don't scale to a statistically reliable sample size.  

The new patent (US 9,008,416), titled "Collection of machine learning training data for expression recognition," covers Emotient's proprietary method for automatically crowdsourcing, analyzing and classifying images of facial expressions quickly and at low cost. This innovation is critical to developing high accuracy emotion analytics services, as the quantity and quality of training data is a key driver of performance, particularly across diverse, real-world applications.

"This patent fortifies our already strong IP position in this space," said Ken Denman, Emotient CEO. "From adding a new 'vital sign' to medical diagnosis to creating advertising that emotionally connects with customers or developing classroom experiences that truly engage students, emotion analytics is poised to drive a revolution in numerous industries."

Training data is used in the fields of artificial intelligence and machine learning to quickly and accurately discover predictive relationships. "It takes a trained human expert about five minutes to accurately evaluate a single face in an image. For this reason, automating how the system learns to accurately recognize facial expressions and emotions is critical to the entire field of emotion analytics," said Javier Movellan, founder and lead researcher, Emotient. "We use crowdsourcing methods to collect and label more than 100,000 images of faces per day to continuously train and improve our automatic expression recognition systems. This patent covers Emotient's program for quickly converting human intelligence into artificial intelligence."

This patent is the second granted in a portfolio of many patents pending which protect Emotient's scientific breakthroughs in this space. It follows an expansive patent granted August 2014 (US 8,798,374 B2) for which Emotient owns the exclusive license rights. That patent covers technology ranging from face detection and primary emotions to detection of facial muscle movements.

About Emotient – Automated Facial Expression Analysis
Emotient, Inc., is the leading authority in facial expression analysis. Emotient software translates facial expressions into actionable information that helps companies make better decisions based on audience response to media, products and experiences. Emotient facial expression technology is currently available through the Emotient Analytics web service, the Emotient AdPanel service or the Emotient API. Automated emotion measurement can generate insights that increase revenue across many industries including advertising, media and entertainment, consumer goods, retail, enterprise sales, healthcare and education.

Emotient was founded by a team of six PhDs from the University of California, San Diego, who are the foremost experts in the application of machine learning, computer vision and cognitive science to facial behavioral analysis. Its proprietary, patented technology sets the industry standard for accuracy and real-time delivery of facial expression data and analysis. Emotient was named a Gartner 2014 Cool Vendor and was recognized as a 2015 TiE Top 50 Startup.

For more information on Emotient, please visit www.emotient.com.


SOURCE Emotient