Demo Traction: This start-up can read your poker face

Facial recognition tech can even tell how a jury will judge a defendant

img 3164

Emotient, which uses video streams for emotion detection and sentiment analysis based on facial expressions, recognizes anger.

Credit: IDG

BOSTON -- The cameras inside a McDonalds recently captured the facial expressions of customers and employees. Every tick, twitch, smile or frown was fed into a computer algorithm.

The algorithm, from start-up Emotient, determined how everybody was feeling about the service: Employees who were stressed out during the lunch hour rush transmitted their stress to customers. Customers were also more unhappy while waiting for their order than when they were waiting to order in the first place.

Emotient, whose software uses video streams for emotion detection and sentiment analysis based on facial expressions, is targeting retail, legal and advertising firms -- any market where reading people's emotions can give them an edge.

The company was among dozens being showcased today at the DEMO Traction Enterprise here.

Emotient CEO Ken Denman said the facial recognition software is faster and less expensive than focus groups, which are also highly inaccurate.

img 3166 IDG

The face of contempt.

"No one tells the truth for a myriad of reasons," Denman said. "Fundamentally, we're disrupting the model for pre-testing in media and ads and...other areas like legal and politics where there are high-value messages that people want to deliver and they want to test those messages...and the stakes are high," Denman said.

Emotient's software keys in on seven base emotions, plus frustration and confusion. The emotions include: joy, sadness, anger, fear, contempt, surprise and disgust.

"We're over 95% accurate across all the major detectors," Denman said.

Emotient's emotion detection software has even been tested in courtrooms, using video of juries to determine how the jurors feel about a closing argument.

Much of the product pre-testing with Emotient's analysis tools is delivered to a demographically selected audience using an online survey platform, which is integrated with RESTful APIs. A group of people who've agreed to be surveyed and recorded login and watch movie trailers, commercials and ad campaigns, for example.

img 3163 IDG

Emotient, whose software uses video streams for emotion detection and sentiment analysis based on facial expressions, recognizes "joy."

Their emotional responses are then logged and sent to the vendor, which can use the results to tailor their products, marketing or ad campaigns.

Emotient's software records at 30 frames per second, so no facial expression, no matter how subtle is missed. There's no way to cover those emotions up with a poker face, according to Marian Bartlett, founder and lead scientists at Emotient.

"There are two pathways in the brain that drive facial movement. You can move your face deliberately...and then there's another pathway... the emotional center," Denman said. "If you're trying to have a poker face..., often what you get are micro expressions.

"The average person doesn't see the micro expressions," she added.

Another presenter at DEMO, Jeremy Allaire, wants digital money to be as easy and free to transfer as an email. Allaire is CEO of Bitcoin start-up Circle Internet Financial, a Bitcoin wallet service, meaning it offers methods to buy and sell the digital currency.

It's being well received by venture capitalists. In the two years since its launch, it has already raised $76 million in three series of funding.

One problem with Bitcoins today is that there are few user-friendly options to use them for making a payment; that, Allaire said, is stifling adoption.

Circle provides mobile apps to enable greater ease-of-use in online and in-person payments, along with enhanced security and privacy for customers.

It's also free to make instant, digital money transfers. Because it's free to send text and photos over the Internet, Allaire said the transfer of money should be free as well.

"Most [online payment systems] are built on the idea that they're going to generate revenue from payments. We don't think there's a business for charging fees for sending money," Allaire said. "We think a dollar should be able to be transmitted as a dollar."

Circle's service is not new. Several other companies also use Bitcoin as part of a payment system, including BitPay, Braintree and Coinbase and Stripe. The latter servcies are used for transactions with eBay and PayPal.

Circle is focused on the consumer market, leaving business-to-business transactions to financial services companies. Circle has set up operations in the U.S., Europe and  China.

Circle is built on block chain-based technology, through which blocks of data are secured with a cryptographic hash. The hash is used to verify the data contained within the message, and any attempt to alter the data creates a new hash that wouldn't match the existing hash on record.

While Allaire didn't say how his company would eventually make money, he did say "things will get interesting" when there's wide adoption.

This story, "Demo Traction: This start-up can read your poker face" was originally published by Computerworld.