Three perspectives on 'The Dress'

Three perspectives on 'The Dress'
Credit: Reproduced with permission from Cecilia Bleasdale.

When you look at this photograph, what colors are the dress? Some see blue and black stripes, others see white and gold stripes. This striking variation took the internet by storm in February; now Current Biology is publishing three short papers on why the image is seen differently by different observers, and what this tells us about the complicated workings of color perception.

Individual differences in color perception uncovered by "The Dress"

For neuroscientists like Bevil Conway, "The Dress" phenomenon marked the greatest extent of individual differences in ever documented. It's long been known that certain optical illusions can cause us to see two different shapes in the same image (e.g., a face or a vase), but what makes "The Dress" photograph so mind-blowing is that it's the first time a single image could be seen by different people as wholly different .

"It caught fire because it was a case in which color wasn't doing what we expect," says Conway, who teaches at Wellesley College and the Massachusetts Institute of Technology. However, the #whiteandgold versus #blackandblue debate on wasn't scientific proof as to how different we each perceived "The Dress." To find out, Conway and his team designed an experiment in which they asked people to identify the colors they saw on "The Dress" from a full palette.

In a survey of 1,400 individuals, with over 300 who had never seen "The Dress" before, Conway and his team found impressive individual differences in color perception; they also found, surprisingly, that people fall into one of three camps corresponding to the main groups identified by social media: a blue/black camp, a white/gold camp, and a smaller blue/brown contingent.

"It could have been the case that you had a continuum of perceived colors, but if you plot the colors people picked, you see two main clumps falling into the two categories for what words people used to describe the colors of 'The Dress,'" says Conway. "This shows that the perception of the dress is variously stable. By studying the pair of colors in 'The Dress,' we can answer the age-old question: do you see colors the way that I see them? And the answer is sometimes 'no.'"

Another finding from the survey was that perception differed by age and sex. Older people and women were more likely to report seeing "The Dress" as white and gold, while younger people were more likely to say that it was black and blue.

Conway believes that these differences in perception may correspond to the type of light that individuals' brains expect to be in their environment. For example, people who perceive "The Dress" as white and gold may have just been exposed to natural daylight, while those who saw a black and blue garment may spend most of their time surrounded by artificial light sources. The brains of those who saw a brown and blue dress are likely used to something in between.

"The big open question is what causes these differences in the population," Conway says. "One framework for understanding why you get these variations is to consider how light is contaminated by outside illumination, such as a blue sky or incandescent light. Your visual system has to decide whether it gets rid of shorter, bluer wavelengths of light or the longer, redder wavelengths, and that decision may change how you see 'The Dress.'"

Current Biology, Lafer-Sousa and Hermann et al.: "Striking individual differences in color perception uncovered by 'The Dress' photograph" http://dx.doi.org/10.1016/j.cub.2015.04.053

'This never would have happened with a red dress'
The original photo which puzzled social network users. Credit: swiked/Tumblr (Reproduced with permission from Cecilia Bleasdale)

The many colors of "The Dress"

In the days after "The Dress" was posted online, a group led by psychologist Karl Gegenfurtner at Giessen University in Germany asked 15 people to view the photograph on a well-calibrated color screen under controlled lighting. The participants then had to adjust the color of a disc to correspond to the colors they saw in the photograph. For the lighter stripe, participants reported seeing a continuous range of shades from light blue to dark blue, rather than white and blue, the two dominant colors reported so far.

"The question should thus not be whether the dress is blue or white, but whether it is light blue or dark blue," write Gegenfurtner and his co-authors. "Despite the continuous choice of matching colors, observers are consistent in calling the dress 'white' when their match lies above a certain brightness and 'blue' when it lies below."

Gegenfurtner's team also found that all of the colors observed in "The Dress" correspond very closely to those found in daylight, adding support to the theory that how the eye interprets natural sunlight is what triggered #Dressgate 2015.

Current Biology, Gegenfurtner,et al.: "The many colors of the dress" http://dx.doi.org/10.1016/j.cub.2015.04.043

The special ambiguity of blue

Would "The Dress" have gone viral had it been #greenandblack or #orangeandblack? Not likely, argues cognitive scientist Michael Webster at the University of Nevada, Reno. He believes that the photograph is part of a growing body of evidence showing that the human eye is more likely to confuse blue objects with blue lighting.

For example, if you stare at a gray object and make the gray increasingly yellow or blue, then you're more likely to see the object as yellow than as blue. This difference likely comes from how the eye evolved in the presence of natural lighting from the sun and the sky.

To test this, Webster and his research team surveyed 87 college students on what color they found the light-blue stripes of "The Dress" to be. The participants were split about fifty-fifty between white and blue. The researchers then inverted the image of the dress so that the black stripes appeared blue and the blue stripes appeared gold. Of those surveyed, nearly 95% said that the stripes were yellow or gold.

"We discovered a novel property of color perception and constancy, involving how we experience shades of blue versus yellow," write the authors. "We found that surfaces are much more likely to be perceived as white or gray when their color is varied along bluish directions, compared to equivalent variations along yellowish (or reddish or greenish) directions."

Current Biology, Winkler et al.: "Asymmetries in blue-yellow color perception and in the color" http://dx.doi.org/10.1016/j.cub.2015.05.004


Explore further

Black/blue or white/gold? Dress debate goes viral

Journal information: Current Biology

Provided by Cell Press
Citation: Three perspectives on 'The Dress' (2015, May 14) retrieved 20 August 2019 from https://medicalxpress.com/news/2015-05-perspectives.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
67 shares

Feedback to editors

User comments

May 14, 2015
asked 15 people to view the photograph on a well-calibrated color screen under controlled lighting. The participants then had to adjust the color of a disc to correspond to the colors they saw in the photograph. For the lighter stripe, participants reported seeing a continuous range of shades from light blue to dark blue, rather than white and blue, the two dominant colors reported so far.


That rather supports the idea that people with cheap monitors see white and gold, and people with better more expensive monitors see black and blue.

Reason being that cheap TN panels distort color towards red/yellow while displaying black as grey, which turns blue into white and black into gold - while more expensive IPS derivatives tend to distort more towards black and purple. These are widely reported properties of the different technologies.

Under color-calibrated conditions, people report seeing blue because the monitors actually display blue.

May 14, 2015
That rather supports the idea that people with cheap monitors see white and gold, and people with better more expensive monitors see black and blue.


That might be true. With my own laptop, I can see all three variants by changing the angle (height) at which I look at the screen. It also worked with two other persons.

May 14, 2015
I'm seeing a dim blue-white and a kind of ugly gold-green.

Controversy!

May 14, 2015
Well i'm getting a scale dependence - smaller pics of the dress i see as white / gold, but larger ones black and blue... with all other factors being equal (same image, monitors etc.) The lower image on this page i see as white & gold, but the zoomed in one above it seems black and blue.

I also tried Googling other copies of it and again, perceived the thumbnails as white / gold but when clicking on them the full-size image appeared black / blue.

Anyone else getting this size dependence effect?

May 14, 2015

I also tried Googling other copies of it and again, perceived the thumbnails as white / gold but when clicking on them the full-size image appeared black / blue.

Anyone else getting this size dependence effect?


The individual pixels as they are displayed are not linear in their brightness values, but corrected according to a "gamma" factor. When an algorithm resamples images, it needs to take this into account because it can't just average the values of two adjacent pixels and arrive at the correct perceptual value.

Hence, the naive rescaling algorithm messes up the colors.

There's also differences in whether the image contains an ICC profile, and whether the browser corrects the colors according to the profile. The thumbnails aren't color corrected and don't contain a profile, so the image gets rendered with false colors. Likewise, a particular web browser may ignore the color profile entirely and render the original image wrong.

May 14, 2015
PNG files contain gamma correction information as well, which different browsers treat differently. Unless the file is specifically stripped of all extra data, the browser may attempt to "correct" the color values before displaying them.

The problem is that Firefox or Chrome or Opera will do it differently from Internet Explorer or Safari, so the same picture looks slightly different on different platforms.

So there's actually several fudge factors in this problem: what kind of monitors people have - are they calibrated, not calibrated, incorrectly calibrated? What browser the user is having? Which copy of the image they're watching, who's edited it, how is it edited...


May 14, 2015
Odd. I saw that pic earlier today and it was blue, but now its yellow. Did you change it?
edit: if you changed the pic thats a cheat.

May 15, 2015
"The original photo which puzzled social network users. Credit: swiked/Tumblr (Reproduced with permission from Cecilia Bleasdale)"

This photo is NOT the original photo, nor is it even a decent reproduction of it.

May 15, 2015
@Eikka - nice attempt at explanations however i'm sure the variable factor is me... tonight i now see the lower picture as blue / black too. I've been trying to flip it back to white / gold, the same way you can sometimes switch your perception of other optical illusions (ie. a vase / two faces in profile), but with this colour constancy illusion i can't manage it...

My guess is that rather than identifying hard-wired differences between distinct groups of people, the illusion is entirely circumstantial. There's probably a way to prime subjects to experience one or other outcome using opponent processing principles or somesuch..

May 16, 2015
As Leibniz argued, the Lockean premise that perception is the basis for scientific inquiry is false. http://www.gutenb...57-h.htm

May 18, 2015
tonight i now see the lower picture as blue / black too.


Daytime natural light tends to overpower computer displays in brightness and your eyes and brain adjust to the different color temperature (5000-10000 K) instead of night-time room lighting (3000-4000 K)

A computer monitor is usually calibrated to 6500 K so it ends up looking yellower when the light in your room comes from the blue sky through a window. They're actually much too blue for regular room lighting, but the eye follows the brightest source of light so staring at the monitor for 5 minutes changes your perception.


May 18, 2015
I've been trying to flip it back to white / gold, the same way you can sometimes switch your perception of other optical illusions (ie. a vase / two faces in profile), but with this colour constancy illusion i can't manage it...


Try this: Set your computer by a window and use it for 15 minutes with a background view of the blue sky.

Then take a desk lamp with an ordinary incandecent bulb or equivalent in a white reflector. Close one eye, point the desk lamp at your other eye for a minute or so, and then try looking at the monitor with one eye and then the other.

The white on the monitor switches between blue/green and yellow/red, and the effect lasts for several minutes.


Please sign in to add a comment. Registration is free, and takes less than a minute. Read more