How does AI make Cinematic Photos in Google Photos?

By Andrew Romero

Google Photos has a few neat automated features called "creations."  One of these creations is Cinematic photos that turn your images into movement-filled animations. These quick animations mimic scenes you might find in a movie, hence the name "Cinematic photos."

Here's how it works...

According to Google, it all has to do with machine learning. First, AI takes a stab at developing the photo into different layers. It determines layers by looking for surface information such as what's in focus and out of focus.

Google claims that no depth data is needed whatsoever to create these layers. Google's AI is just that good.

The next step is filling in visual gaps between the layers. The idea is that when the layers move, no visible separations are in view. Google Photos sort of airbrushes these areas to maintain continuity.

Once those layers are created and gaps filled in, Google Photos' AI can then animate a virtual camera, moving the layers and point of view to give the illusion of a 3D image. 

The final product you're left with is a stunning, Cinematic photo that couldn't be more personal. Google Photos will generally create these out of pictures that include the following: - A subject such as a person - A decent amount of natural bokeh and depth of field

Unfortunately, you can't make these yourself like you can other creations in Google Photos, though this might change in the future. To keep up with Cinematic photos and other Google news, head to 9to5Google and check out some of our other stories.