In creating Real Tone processing for the Pixel 6, Google sought to give people of color a camera treatment that wasn’t based on close to two centuries of White-centered photography. But it’s not enough for just one device to have this kind of approach to imaging, so at Google I/O 2022, the company is pushing it out to Google Photos and is open-sourcing the engine behind the magic.
Later this month, Real Tone will be available to Google Photos users as filters — designed by “a diverse range of renowned image makers” known for accuracy — that they can apply to all sorts of pictures they’ve taken across all sorts of skin tones. Those who have been looking for hues to complement their warm or cool appearance should be excited to test these out.
“Monk Skin Tone Scale” by Dr. Ellis Monk, made available by Google LLC under the Creative Commons Attribution 4.0 International license.
In addition, Google has made the Monk Skin Tone Scale, which the company has used to evaluate its color adjustment decisions against, open source. The scale is comprised of ten tones, is named after Dr. Ellis Monk, an associate professor of sociology at Harvard University, and can be applied to machine learning models for computer vision projects of all sorts. Color values are available in HEX, RGB, HSL, CIE-LAB, and CIE-LCH.
The company is pushing researchers to follow a list of recommended practices while they conduct their experiments including making sure that subjects reflect the totality of the scale, that intersectional factors be considered when analyzing data, and that skin tone does not equate with race.
The MST Scale will be updated over time and those who use it don’t necessarily have to adhere to the strict color values listed, but the main idea here is to start from a more accurate basis for understanding what we perceive about skin tone and what relates with it.