INNOVATION

Can Facial Recognition Really Tell If a Kid Is Learning in Class?

Inventors of software called EngageSense say you can tell if kids are engaged in class by analyzing their eye movements

November 1, 2013

Can cameras read what’s going on in a second grader’s mind? Photo courtesy of EngageSense

All of us have had a teacher who had eyes in the back of his or her head. Even while facing the blackboard, they saw everything—every note being passed, every answer being copied, every face being made.

Or at least it seemed that way. All they really had to do was guess right a few times about what was going on behind their backs and, well, that is how classroom legends are made.

But what if you took all the guessing out of the picture? What if cameras focused on every kid in the class? That’s what a New York company named SensorStar Labs has in mind, although the point would not be to catch miscreants, but rather to help teachers determine when they’ve lost the class.

Face time

Here’s how it would work. Using facial recognition software called EngageSense, computers would apply algorithms to what the cameras have recorded during a lecture or discussion to interpret how engaged the students have been. Were the kids’ eyes focused on the teacher? Or were they looking everywhere but the front of the class? Were they smiling or frowning? Or did they just seem confused? Or bored?

Teachers would be provided a report that, based on facial analysis, would tell them when student interest was highest or lowest. Says SensorStar co-founder Sean Montgomery, himself a former teacher: “By looking at maybe just a couple of high points and a couple of low points, you get enough takeaway. The next day you can try to do more of the good stuff and less of the less-good stuff.”

No doubt some parents are going to have a lot of questions about what happens to all that video of their kids’ faces. But Montgomery is confident that most will agree to let their children be videotaped when they see how much it helps teachers polish their skills.

He’s convinced that in five years, teachers all over the country will be using it. First, though, he has to prove that the SensorStar algorithms can truly interpret the workings of young minds based simply on eye movement and facial expression.

Small measures

That, of course, assumes teachers will jump right on board. Which is hardly a sure thing, given the response last year to a report that the Bill and Melinda Gates Foundation is helping to fund the development of sensor bracelets that could, in theory at least, track a student’s engagement level.

The wrist devices are designed to send a small current across the skin and then measure subtle changes in electrical charges as the nervous system responds to stimuli. These bracelets have been used in tests to gauge how consumers respond to advertising, and the thinking goes that if they can tell you how excited someone gets while watching a car ad, they can give you a sense of how jazzed a kid can get about fractions. (Or not.)

Not so fast, snapped skeptics. They were quick to point out that just because a second grader is excited doesn’t mean he or she is learning something. And while the bracelets’ boosters argue that their purpose is to help teachers, critics say that no one should be surprised if the sensors eventually are used to evaluate them. Some teachers suggested that they might have to work random screams into their lesson plans to keep the excitement level high.

Ultimately, it comes down to whether, like Bill Gates, you believe that accumulating and analyzing data from classroom behavior is the key to applying science to the learning process. Or, if you think that teaching is more art than science, and that the connection between teachers and students is too complex and nuanced to be measured through a collection of data points.

Who’s your data?

And you will not eat a salad your first six months in college: More and more colleges are using predictive analysis to give students a good idea of how they’ll fare in a class before they even sign up for it. By using data from a student’s own academic performance and from others who have already taken the class, advisers can predict with increasing accuracy how likely it is that a particular student will succeed or fail.
Please like this investment: Last week Facebook founder Mark Zuckerberg made his first investment in a startup company—he joined a team of investors putting $4 million in seed money behind a Massachusetts company named Panorama Education. It crunches data from surveys it does for schools from K to 12, ranging from subjects such as why some promising students end up failing to why bullying is particularly prominent among ninth grade boys.
Acing the tests: A smartphone app called Quick Key has an optical scanner that can quickly grade SAT-style bubble answer sheets. Then it uploads the results to teachers’ electronic grade books and analyzes the data.
Apple-picking time: Earlier this week, Apple CEO Tim Cook announced that iPads make up 94 percent of the tablets now used in schools. The company’s sales have slowed in the consumer market, so it’s been making a big push into education by offering discounts for bulk purchases.
And they probably drew outside the lines: A new study from Michigan State University found that people who were involved in artistic activities while they were in school tended to be more innovative when they grew up—specifically that they were more likely to generate patents and launch businesses as adults.

Video bonus: Bill Gates offers his take on how he thinks teachers should be given feedback.

Video bonus bonus: Here’s a different twist on facial recognition in the classroom.