Friday, May 19, 2017

Stripey things, zebras, and the uncanny valley

Stripes and zebras
Stripes can mess with your brain. A recent study from research in the Netherlands and the U.S. suggests that “looking at intensely stripey things causes an increase in gamma oscillations in the brain” which can be linked to headaches and seizures. Many people just find stripes weird, but some experience very real visual distortions. What is even more interesting is that these effects are more likely to be caused by human-created stripes such as venetian blinds, rather than natural stripes, like those found on zebras. Researchers found that distorting the lines slightly or blurring their edges caused the oscillations to die down. And vertical stripes are not as disturbing as horizontal ones. “It seems that our brains are not designed to cope with such extreme regularity, as it doesn’t occur in nature.”

The Uncanny Valley
Natural realism and artificial realism are also the basis for ongoing research into an odd property of computer generation called “uncanny valley”. The closer the images get to total realism the more disturbing they seem to become. Japanese robotics engineer Mashahiro Mori coined the term in a paper he wrote in 1970 titled The Uncanny Valley. He proposed that we will accept a synthetic human that looks and moves realistically, but only up to a point. Once the resemblance comes close to, but not close enough to reality, we become more and more disturbed by slight anomalies. Mori’s theory made its way into computer animation. Stylized cartoons engender empathy but pseudo-human characteristics can easily go awry.

"Uncanny valley really does relate to painting! The closer the work gets to being
realistic the more cognitive dissonance is triggered. This causes a person to feel really uncomfortable, so the mind jettisons whatever is causing the dissonance." 
 Kathryn Fisher, Artist

When the first computer-generated elements began turning up in Hollywood films, technicians were capable of making things like dinosaurs, metal men, and spaceships, but creating a realistic human, with all its variety and subtle changes, seemed unattainable. The outward appearance of a human or human face could be created, but all the variables present in reality, especially having to do with subtle movement, were more difficult to achieve. Even slight imperfections in humans can create unsettling reactions in viewers. The closer to reality an animation becomes, the more likely it is to create cognitive dissonance and a sense of discomfort and conflict in the viewer.

We pay attention when something is changed,
 or different, or just seems weird.

Seeing is dependent on noticing and we notice only when we look for something. We cannot notice everything, but we do pay attention when something is changed, or different, or just seems weird. The process of visual observation is a complex one. We notice the most obvious information and tend to overlook all the nuanced information that actually underlies our perception.


Leonardo da Vinci’s painting Mona Lisa is probably the most iconic painting in the world. We also know how this painting “looks”. But art historian E. H. Gombrich pointed out how difficult it is to look at this painting with fresh eyes. He urged viewers to look anew, to try and forget what we think we know and focus on what we truly see. “She really seems to change before our eyes and to look a little different every time we come back to her.”

Neurobiologist Margaret Livingstone, author of Art and Vision, did just that. She looked anew at the painting and noticed the expression on Mona Lisa’s face was dependent on the discrepancy between our peripheral and central vision systems. The center of our gaze is optimized for small, detailed things, while our peripheral vision has a lower resolution and is better at big “blurry” things. We are usually not aware of this difference because we are constantly moving our eyes around, and we do not notice that our peripheral vision (blurry) can be just as important as our central vision (detail). If you move your eyes around the painting, her expression appears to change. Look directly at her mouth and she appears to smile less than when you’re staring at her eyes. Our peripheral vision picks up the slight shading around the mouth which gives the impression of a smile. When your gaze falls on the background or on her hands, this effect can be even more pronounced. 
(Livingstone notes this observation is more apparent when viewing the original painting instead of a reproduction.)

More stripes, more zebras
Dazzle camouflage was a type of ship camouflage used extensively in World War I and to a lesser extent in World War II. British marine artist Norman Wilkinson is usually credited as being the father of dazzle camouflage but this is not entirely accurate. The idea was initially proposed by the British zoologist John Graham Kerr. In writing to Winston Churchill in 1914 he explained the goal was to confuse, not to conceal, by disrupting a ship's outline. Kerr made the comparison to the patterns on land animals such as the zebra and suggested a similar pattern but with the use of countershading to also offer a measure of invisibility.


American artist Abbott Handerson Thayer wrote to Churchill in 1915 and suggested disruptive coloration and countershading based on his 1909 book Concealing-Coloration in the Animal Kingdom. Neither Thayer nor Kerr were able to win over the Admiralty, but along came Norman Wilkinson, a marine artist and Royal Naval Volunteer Reserve officer. He advocated "masses of strongly contrasted colour" to confuse the enemy about a ship's size, speed and heading. He also said the effect was not to conceal, but to cause the enemy to take up a poor firing position.  Kerr, whose proposal was based on years of study, lost out to the more socially connected Wilkerson. Later, Kerr was asked if he had, in fact, invented dazzle camouflage and he replied by saying "this principle was, of course, invented by nature."
Thanks to Artist Amber Blazina for this interesting tip on dazzle camouflage.


And for those of you who have read this far, here's a diagrammed selection from John Singer Sargent's painting The Daughters of Edward Boit. Any figure or any "thing" we paint needs a variety of edges, most notably from side to side. Too many similar or hard edges, especially on a horizontal plane, can negate our attempts at creating the illusion of three dimensions. I have illustrated a few obvious points of reference and also included some diagonals to illustrate the flow of information and the variety of edges.
Suggestion: Try squinting at the image to see value contrast more easily, and keep in mind edge information can be enhanced by value similarities or value differences between the object and the area around the object.