Imagine that you wanted to train a facial recognition algorithm on age-related characteristics and, more specifically, on age progression (e.g., how people are likely to look as they get older). Ideally, you’d want a broad and rigorous dataset with lots of people’s pictures. It would help if you knew they were taken a fixed number of years apart—say, 10 years.
Sure, you could mine Facebook for profile pictures and look at posting dates or EXIF data. But that whole set of profile pictures could end up generating a lot of useless noise. People don’t reliably upload pictures in chronological order, and it’s not uncommon for users to post pictures of something other than themselves as a profile picture. A quick glance through my Facebook friends’ profile pictures shows a friend’s dog who just died, several cartoons, word images, abstract patterns, and more.
In other words, it would help if you had a clean, simple, helpfully labeled set of then-and-now photos.
What’s more, for the profile pictures on Facebook, the photo posting date wouldn’t necessarily match the date the picture was taken. Even the EXIF metadata on the photo wouldn’t always be reliable for assessing that date.
Why? People could have scanned offline photos. They might have uploaded pictures multiple times over years. Some people resort to uploading screenshots of pictures found elsewhere online. Some platforms strip EXIF data for privacy.
Through the Facebook meme, most people have been helpfully adding that context back in (“me in 2008 and me in 2018”) as well as further info, in many cases, about where and how the pic was taken (“2008 at University of Whatever, taken by Joe; 2018 visiting New City for this year’s such-and-such event”).
In other words, thanks to this meme, there’s now a very large dataset of carefully curated photos of people from roughly 10 years ago and now.
Read the full article @ Wired