-
It never forgets a face.
August 08, 2016It Never Forgets a Face
You might be good at remembering faces, but statistically, computers are now better at it than you.
The human brain evolved over millions of years to focus on people’s faces and match them to stored memories. Cameras and computers have to be taught to look for those characteristics and process the data. While they could recognize you, there were plenty of glitches.It Never Forgets a Face
You might be good at remembering faces, but statistically, computers are now better at it than you.
The human brain evolved over millions of years to focus on people’s faces and match them to stored memories. Cameras and computers have to be taught to look for those characteristics and process the data. While they could recognize you, there were plenty of glitches.It Never Forgets a Face
You might be good at remembering faces, but statistically, computers are now better at it than you.
The human brain evolved over millions of years to focus on people’s faces and match them to stored memories. Cameras and computers have to be taught to look for those characteristics and process the data. While they could recognize you, there were plenty of glitches.Faces are complex, with characteristics that constantly change. Temporary facial expressions plus changes in hair, make-up and blemishes made matching difficult. To complicate things, cameras capture different angles and lighting, making it even more confusing for software.
Faces are complex, with characteristics that constantly change. Temporary facial expressions plus changes in hair, make-up and blemishes made matching difficult. To complicate things, cameras capture different angles and lighting, making it even more confusing for software.
Faces are complex, with characteristics that constantly change. Temporary facial expressions plus changes in hair, make-up and blemishes made matching difficult. To complicate things, cameras capture different angles and lighting, making it even more confusing for software.
Technology has evolved to the point that facial changes are no longer the obstacle they once were, according to Jim Howard and James Snyder, both Ph.D. educated 3M software engineers. 3M’s software relies on the computer’s ability to “learn” the best landmarks. The software is biologically inspired, James explains. “Its techniques were originally modeled on our brain processes,” he says.
Technology has evolved to the point that facial changes are no longer the obstacle they once were, according to Jim Howard and James Snyder, both Ph.D. educated 3M software engineers. 3M’s software relies on the computer’s ability to “learn” the best landmarks. The software is biologically inspired, James explains. “Its techniques were originally modeled on our brain processes,” he says.
Technology has evolved to the point that facial changes are no longer the obstacle they once were, according to Jim Howard and James Snyder, both Ph.D. educated 3M software engineers. 3M’s software relies on the computer’s ability to “learn” the best landmarks. The software is biologically inspired, James explains. “Its techniques were originally modeled on our brain processes,” he says.
“It’s a breakthrough. Traditional recognition was based on certain landmarks, such as points on the face. We have moved away from defined landmarks. Human intuition is not directing the algorithm. It’s learning its own set of features, so it can still estimate features whether someone is smiling or frowning.”JIM HOWARDSOFTWARE ENGINEER, 3M SAFETY & GRAPHICS LAB
The Four Steps of Facial Recognition
Law enforcement agencies can use the facial recognition system to help identify suspects, clear innocent people and identify people who aren’t able to identify themselves due to dementia or amnesia.
The bottom line? Everyone is unique, and technology is quickly figuring that out.1. Find the Face: Background separation, precise location tracking
2. Extracting Features: Essential information, template creation
3. Normalizing the Face: Pose correction, image adjustment, landmarks
4. Matching: Scoring, score fusion
1. Find the Face: Background separation, precise location tracking
2. Extracting Features: Essential information, template creation
3. Normalizing the Face: Pose correction, image adjustment, landmarks
4. Matching: Scoring, score fusion