Facebook added the ability to automatically convert 2D images into 3D images using AI

Last year, Facebook announced a relatively interesting new feature, allowing users to view and post 3D photos. Immediately after its launch, this feature has received the attention of a large number of Facebook users, mostly interesting and appreciated. However, it is limited to use with modern high-end smartphones that own a camera that comes with a portrait mode (remove fonts).

In order to help the 3D image feature reach more users, Facebook has developed a new machine learning technique that can automatically convert any 2D image into 3D image in a short time. Whether it's photos from standard single cameras or even photos that have been around for decades, just upload them to Facebook, and the 3D photo converter will take care of the rest.

With a standard RGB image, Facebook's 3D image processing neural network can estimate the distance from the camera to objects based on each pixel, then create a 3-dimensional background and font removal effect. the most reasonable way. This approach basically helps Facebook can estimate the depth of the 2D image to produce a complete 3D animation without the need for a camera to remove the background.

In fact, the use of artificial intelligence algorithms to build images to remove fonts is not new. Google has used the same technique for portrait mode on its single Pixel camera phones. However, the advantage of Facebook is that it allows users to convert even photos taken before, unlimited. In addition, this converter will be integrated directly into the Facebook platform, users can experience it directly without having to download the application or use third-party 3D photo rendering tools as before.

If you own an iPhone 7 or newer or newly launched mid-range Android devices, you can try the 3D photo rendering feature in the Facebook app.