Computers fully comprehending things they see in front of them will be the technical advance that takes us to the next level in our digitized world, experts say.
“The next leap in computing will be in how we interpret images,” said Jem Davies, an ARM executive at an embedded technology conference recently.
“That will be revolutionary,” he said, according to IEEE Spectrum, writing about the processor executive’s talk in the article “Bringing Eyes to the Internet of Things.”
One thinks of objects in this context. But a facet to this camera-based artificial intelligence is that computers will also commonly perform facial recognition. Computers will likely become better at recognizing people than humans. Not only that, but they will be better than us at recognizing emotions through facial expressions.
Facial recognition and school tests
Kids may be the first to experience automated emotion capture.
Facial and eye tracking will be used to augment childrens’ school tests, says Hong Kong-based F.S. Artificial Intelligence Ltd., an organization I met with at a TechCrunch Disrupt startup fair earlier this month.
The organization has developed a way to use an off-the-shelf tablet device with a standard camera to capture feelings expressed by the kids as they answer questions.
Gaze tracking evaluates the emotion of the individual while capturing the test responses and lets teachers analyze how receptive the kids are to the questionnaire. The company says it helps with mental health alerts and the effectiveness of the test design. For example, it can see if the kids get distracted because the questions are too easy.
I had a chance to try the technology at the show. They used arithmetic on me, and I reckon the results pretty accurately reflected how I felt: I registered surprise when I saw the first (slightly difficult) question and happiness when I got the second (easy) question correct.
The eye and facial tracking and big data-style analysis can also be used in advertising predictions and medical diagnosis, researchers say.
“Poker players supposedly manipulate their expressions to mislead,” says the Financial Times, writing about how human feelings will be detected by computers in the future (paywall).
Those days may be limited by these new methods. The newspaper says it has seen demonstrations of facial recognition software that can detect the difference between a put-on smile and a real smile while the subject watches a TV show.
The company producing the software, CrowdEmotion, uses its artificial intelligence algorithm in the television industry to gauge reactions to programming before programs are aired.
Facial recognition system better than humans
And Chinese search giant Baidu says its face recognition system is already better at recognizing faces than humans, according to Daily Mail, which wrote about a Chinese tourist town’s facial recognition efforts to track whether tourists have paid to wander around or not.
“The system uses cameras to spot people as they approach the town entry and checks these against a database of registered visitors within a few seconds,” the publication says. It’s apparently 99.77 percent accurate and doesn’t need intrusive gates.
“There’s a dichotomy between what people say and what they feel,” says Jinghan Ong, of CrowdEmotion says in the Financial Times article. “In focus groups, they will say they hate something, likeKeeping Up with the Kardashians, but their face shows that they are enjoying it.”