Google Photos uses machine learning to make your images look 3D
Smartphone cameras have become increasingly sophisticated with portrait features, night modes, and more. But how about making your images look more cinematic? Google Photos has announced a new feature that will do just that.
Google is calling the feature “cinematic photos.” Essentially, the feature will “use machine learning to predict an image’s depth and produce a 3D representation of the scene—even if the original image doesn’t include depth information from the camera.” Google will then apply a small animation to the photo, giving it a smooth panning effect.
The result Google shared gives the photo much more life and depth. And it gives the impression that a much more expensive camera was used to take the photo, when in reality it’s all thanks to Google’s sophisticated machine learning technology.
Google Photos will automatically create cinematic photos for you as long as your app is updated. When a cinematic photo is created, it’ll show up in your recent highlights at the top of your photo grid. You can then share the result with friends and family. Keep in mind that every cinematic photo is shared as a video.
As cool as the new feature looks, it gives me slight pause. With Google Photos getting rid of free unlimited storage next year, I wonder how much space each cinematic photo will take up. Regular smartphone images hardly take up any space, but turning many of them into short videos could add up.
In addition to cinematic photos, Google also announced Memories will now regularly surface photos of the most important people in your life and also your favorite activities, such as baking or hiking. Google’s algorithms will apparently determine your favorite things based on the photos you upload.
You’ll start seeing cinematic photos in Google Photos over the next month, so keep an eye out.
Google Photos (Free, Google Play) →
The post Google Photos uses machine learning to make your images look 3D appeared first on xda-developers.
from xda-developers https://ift.tt/3r2s6fF
via IFTTT
Aucun commentaire: