Image-Melding, from Google Photos

Editor’s Note: This article previously appeared in a different format as part of The Atlantic’s Notes section, retired in 2021.

Recently I recommended that you check out Google Photos if you have not done so already. Like Gmail, it’s a way to store huge quantities of digital material and leave its management to someone else. (I promise, later we’ll get into the privacy tradeoffs involved.) And much more than Gmail, it offers big-data tools that can arrange and transform your information/photos in ways difficult or impossible to do by yourself.

For instance: I mentioned that Google Photos had, on its own, merged three smart-phone snapshots of a scene at Oxford into one panorama view. Several people wrote in to say: Let’s see the originals! So here goes.

First, in its full-frame entirety, a smartphone snapshot of one side of the entry quad at The Queen’s College, Oxford.

Then two almost-identical shots of the other side, both in full frame. First this:

Then this:

The point is that without my doing anything more than saving all three shots to a Google drive, the system recognized them as overlapping parts of a whole and stitched them together into a high-rez, level-horizon, panorama version, looking like this (and at larger scale here):

Even when zooming in on the composite shot as far as possible, I still can’t find a pixellated boundary where the shots were brought together.

We all say in our blase way: Yeah yeah there’s increasing power of big-data systems. At least for me, seeing how it worked on my own information dramatized these effects. To be clear, this was with three quick, casual phone-shots taken over a few-second span. The result isn’t anything fancy, but it’s different from what I could have done myself.

And, as I say, we’ll get to the surveillance-state ramifications soon.

***

Actually, why not now. Here’s one reader response:

In the vein of the glass being neither half-full nor half-empty, but having a leak, that [big-data] magic means just as much that a computer can figure-out where you have been and when based on the photos you take, making it that much easier for a human being with access to that computer to know where you have gone and when.

The photo taker providing in essence CCTV of their movements.  At least though unlike CCTV (and of course all that magical facial recognition) if the photo taker stops taking/posting the photos the intelligence stream stops.