There’s a scientific reason why some human-like faces appear cute, while others definitely do not.
Our brains are wired to recognize faces in everything — clouds, rocks, burnt toast, the moon. We try to make sense of random patterns and often interpret them as faces, and this phenomenon is known as facial pareidolia.
According to a study completed by Canadian and Chinese researchers, this is most likely an evolutionary survival mechanism that came about from the necessity of being able to recognize family, friends and foes. And it is perfectly normal.
To take this phenomena to the next level, Robby Kraft, a science- and math-based artist, decided to use a software called AverageExplorer, recently developed at University of California, Berkeley, that can take thousands of images or photographs and combine them in a way that smooths out all the features, producing an average composition. Kraft wanted to know what would happen if he used a bunch of objects that only looked like faces.
An example of facial pareidolia. Photo credit: Nottsexminer/Flickr (CC BY-SA 2.0)
So how exactly does the software work? First, it has to be able to find and tag features from each object such as eyes, a nose and a mouth. It then has to scale the images, rotate and align them, and finally find a way to blend and shade the colors. Kraft adjusted the software as needed and then applied it to 2,500 objects. The software was only able to identify faces in 100 of the objects, however that was enough to create the ghost-like image seen above.
And it is a very creepy image — something from a horror movie or a nightmare. But just think about it, no human faces were used to create this image — only inanimate objects.
Does this “face” look disturbing to you? If it does, you are not the only one. This image definitely falls under what is called the “uncanny valley” effect — it looks human enough to be recognizable as one, but there is something that makes it appear off-putting and creepy.
This diagram explains the effect:
Photo credit: Smurrayinchester/Wikipedia (CC BY-SA 3.0)
The concept of uncanny valley was first identified by robotic professor Masahiro Mori in 1970. He stated that as a robot, for example, is made to look more human, some observers’ emotional response will become positive and empathetic, until a point is reached and the response quickly turns to revulsion. However, as the robot’s appearance continues to once again look less and less from a human, the emotional response switches back to positive.
For example, the robot on the left is considered cute while the one on the right is a little revolting:
Photo credit: bcogwene/Pixabay and BradBeattie/Wikipedia (CC BY-SA 3.0)
How we interpret faces is all in our heads, while some people recognize faces in almost everything, others do not. Do you find any of these images uncanny?
You may also like: Robots Can Convince People to Share Their Deepest Secrets