This week Facebook quietly rolled out Shared Albums, which lets you co-create massive galleries with your friends. The idea is that all those collective pictures from birthdays, trips, weddings, parties, etc that get spread out between the bunch of you will now have one home. The album’s max photo limit? 10,000 photos. Talk about comprehensive.
It’s a smart feature for Facebook to roll out, especially considering that more than a couple third party apps had solutions out there to fill the void. So why the wait? All Facebook told us was that the update came after a hackathon earlier this year and that it was a popular user request.
Apparently, these researchers knew things were heading in Shared Albums direction – and maybe even have a few upgrades that could improve it. An article titled Analyzing Facebook features to support event detection for photo-based Facebook applications was published in 2012 at a German University and basically hits the sharing albums nail on the head.
“Until now there is no easy way to detect and link photos that are related to the same events, which are usually distributed between between friends and albums. In this work, we introduce an approach that exploits Facebook feature to link photos related to the same event.”
Sound familiar? The article, however, goes on to suggest methods for taking things up a notch:
“In the current situation where the EXIF header of photos is missing in Facebook, we extract visual-based, tagged areas-based, friendship-based and structure-based features. We evaluate each of these features and use the results in our approach. We introduce and evaluate a semi-supervised probabilistic approach that takes into account the evaluation of these features.”
So basically, instead of having to request people get in on the album sharing business and manually upload the photos from whatever fun group activity you guys did not invite me to, this proposed system would analyze data points like what the photos looked like, if they were location tags, and if similar people were in them. It sound as if a shared album would just be auto-created if you and you’re friends all uploaded images from the same event because the service could tell that the place was the same, the people were the same, etc. Given that Facebook is expanding it’s already massive facial recognition database, it certainly seems possible.
And sure, while Facebook pre-determining these elements means things could be hit or miss (we’ve all seen the social network pull a drunken, “Is this you?” face tag request indicating it thinks you’re an elbow), it’s also intriguing. Imagine seeing a photo album of your hiking trip, full of everyone’s photos, the minute those images are uploaded.
Maybe these features will be added to Shared Albums in the future? “In this approach we create a look-up table of the initialization values of our model variables and make it available for other Facebook applications or researchers to use. The evaluation of our approach showed promising results.”