Simple Visual Illusion Dupes Computer

visual line illusions — In the classical four-wing form of this illusion, the top line appears shorter than the bottom line, even though the lines are of equal length. Terminating circles still induce a perceptual effect of line length misjudgment, as will arrows with the line shafts removed.

(Image credit: Astrid Zeman et al., PLOS ONE, doi:10.1371/journal.pone.0056126)

Even computers can get tricked by optical illusions, a new study finds.

Such research may help shed light on how vision works in the brain, and lead to better computer recognition of images, scientists added.

Optical illusions, more properly known as visual illusions, take advantage of how the brain perceives what the eyes tell it in a way that plays a variety of tricks on the mind. For instance, these illusions may cause people to see something that is not there, or not see something that is there, or see an unrealistic portrayal of object, or see one thing as two or more completely different things. By investigating how illusions fool the brain, researchers can learn more about the brain's inner workings

Latest Videos From

Watch full video here:

"Recently, many computer models have tried to imitate how the brain processes visual information because it is so good at it," Zeman said. "We are able to handle all sorts of changes in lighting and background, and we still recognize objects when they have been moved, rotated or deformed. I was curious to see whether copying all of the good aspects of object recognition also has the potential to copy aspects of visual processing that could produce misjudgments."

"What is exciting about these results is imagining what would happen in the future," Zeman told LiveScience. "If we build robots with artificial brains that are modeled off our brains, the implication is that these robots would also see illusions much like we do. By imitating the amazing accuracy, flexibility and robustness that we have in recognizing objects, we could also be copying potential errors in computation that manifest in visual illusions."

The researchers first showed pairs of lines to a computer model of human vision. Each pair had one line that was longer than the other. Each line either had both an arrowhead and an arrow tail or an "X" at both ends. The computer model, named HMAX, had to guess which line was longer, and it was told when it was correct and when it was wrong. In this way, the investigators trained the system to correctly identify what long and short lines look like with 90 percent accuracy.

"We train a biologically plausible model and look at the influence of the images it is exposed to," Zeman said. "If we think of this visual system as something we implant in a robot, this means that we can grow whole bunch of robots up in different environments. Then, once our robots have matured and have learnt to see things, we can then smash their brains open to see what they are thinking. This is something that we can't quite do with humans."

These findings may eliminate a number of potential explanations for the illusion. For example, in the past, scientists had speculated this illusion was caused by human brains misinterpreting arrowheads and arrow tails as depth cues — in modern-day environments, rooms, buildings and roads present boxy scenes with many edges, and so might lead people to unknowingly make predictions regarding depth whenever they run across angles and corners. However, since this computer model was not trained with 3D images, these findings may rule out that idea. [The 10 Greatest Mysteries of the Mind]

"There are so many visual illusions that exist out there, and new ones are coming out all the time," Zeman said. "These illusions bring to light new questions about how we perceive the world and the assumptions we make about the world. Currently there is no existing formal and comprehensive catalog of illusions, so one direction for future development would be to pool together all of this knowledge."

Charles Q. Choi is a contributing writer for Live Science and Space.com. He covers all things human origins and astronomy as well as physics, animals and general science topics. Charles has a Master of Arts degree from the University of Missouri-Columbia, School of Journalism and a Bachelor of Arts degree from the University of South Florida. Charles has visited every continent on Earth, drinking rancid yak butter tea in Lhasa, snorkeling with sea lions in the Galapagos and even climbing an iceberg in Antarctica.