It's been a good year so far for Google Photos, which keeps getting great free upgrades that add neat new features. And another one has just dropped.
The latest upgrade, announced on the official Google Photos Twitter account (opens in new tab), will make a big difference to photos of people – and it's rolling out right now to everybody who uses Google Photos on Android phones, on iPhones or on the Google Photos website.
The upgrade brings new Real Tone filters to your Google Photos filter collection, and the new ones are clearly marked with "Made With Real Tone", as in the screenshot above. If you've found your filters a little unpredictable or disappointing with certain skin tones, the new filters should fix that for you.
What are Google Photos Real Tone filters?
Real Tone filters are based on something called the Monk Skin Tone Scale (opens in new tab), an open source library Google hopes will solve an ongoing problem with computer and smartphone imaging: many imaging apps work brilliantly with white skin but don't cope so well with darker skin tones. One of the most famous examples of that was HP's supposedly racist webcam back on 2009: its face tracking software didn't work for many Black people. The problem turned out to be with its processing in some lighting scenarios rather than any inherent racial bias, but systems such as automatic facial recognition may misidentify (opens in new tab) Black people five to ten times more often than they do white people.
Of course, disappointing photo filter results aren't as big a deal as facial recognition failing to recognise you or thinking you're someone you're not. But Google's photo filters are just a happy spin-off from the more serious project of ensuring that computer vision technology works for all people, not just white people. As Google rightly says, that requires intentional changes to make systems more inclusive from the data they're based on to the algorithms they use – and the more we rely on computers to see and process the world, the more important that inclusivity becomes.
