Facebook Adds New Option to Create 3D Photos From Any 2D Image
2020-02-29 | Since 5 Month
Facebook has added a new element to its 3D photos option, with its system now able to simulate depth in any image, enabling you to convert any static picture into a 3D post.
As explained by Facebook:
"Our 3D Photos feature has [thus far] relied on the dual-lens “portrait mode” capabilities available only in new, higher-end smartphones, so it hasn’t been available on typical mobile devices, which have only a single, rear-facing camera. To bring this new visual format to more people, we have used state-of-the-art machine learning techniques to produce 3D photos from virtually any standard 2D picture. This system infers the 3D structure of any image, whether it is a new shot just taken on an Android or iOS device with a standard single camera, or a decades-old image recently uploaded to a phone or laptop." The process adds to Facebook's 3D image options - Facebook first launched 3D image posts back in 2018, then added the same for Stories last year. But as Facebook notes, up till now, the option has been limited to higher-end devices, utilizing their multi-camera capacity to facilitate 3D depth of field.
Now, through this new process, any image, captured on any device, can be given a similar effect.
To try it out for yourself, you should first update your Facebook app to the latest version. After that, tap on your 'What's on Your Mind?' post creation prompt at the top of your News Feed and scroll through the post options below. You'll see that there's now a '3D photo' option available - tap on that and it will take you through to your Camera Roll where you can select an image,
Select whichever image you want, and Facebook will translate it into 3D (second screenshot above), after which, you'll get a 3D simulated post like this. It's not true 3D, of course, as it's not full depth, but it's a simulation of 3D modeling, which still adds a similar effect, and responds when you tilt your device.
To create this effect, Facebook has used a new process of image recognition and object perception,
"We trained a convolutional neural network (CNN) on millions of pairs of public 3D images and their accompanying depth maps, and leveraged a variety of mobile-optimization techniques previously developed by Facebook AI."
Facebook has provided a full rundown of its process here, which covers the more technical elements of the system. But the basics are that the process can identify theoretical foreground and background elements, based on other 3D image examples, then split them in order to simulate depth.
As a technical achievement, it's pretty amazing, and while it's not full 3D, as noted, it can add an interesting angle (see what I did there) to your Facebook posts, and could help generate more interest.
Of course, from a marketing and promotion standpoint, you're better off creating full 3D images through Facebook's existing 3D post process, as that actually will provide full depth. But maybe there are some older images you could try out, and add a new perspective to.
And Facebook's not done there: "In addition to refining and improving our depth estimation algorithm, we’re working toward enabling high-quality depth estimation for videos taken with mobile devices. Videos pose a noteworthy challenge, since each frame depth must be consistent with the next. But it is also an opportunity to improve performance, since multiple observations of the same objects can provide additional signal for highly accurate depth estimations."
That could be another option in future, creating more immersive video content, which would also lean into the AR and VR shift.