Lumiere, how to use Google AI to create videos from photos

Google moves into the future with confidence, releasing new models and new tools based on Artificial Intelligence. Thanks to the collaboration with the Weizmann Institute of Science and with Tel Aviv University, the Mountain View company is developing Lumiere, a new software with which to create extremely realistic videos starting from photographs and images.

Among the many tools based on generative AI that are populating the web, Lumiere, with its evocative name, promises to take numerous steps forward and attract the attention and curiosity of users. Let’s find out what it is, how it works, what content it allows you to create and what times are expected for release.

Lumiere by Google, how it works

It is called Lumiere and its name is taken directly from the Lumière brothers, inventors of the cinema. On the other hand, the new AI-based system launched by Google wants to bring movement in static imagescreating something spectacular and captivating.

Lumiere is capable of make short videos inserting dynamism into the images provided by the user. The duration of the videos returned by the system is, at the moment, a maximum of five seconds.

Animation can be applied to the entire image, just a part or just small elements. The Directions are provided through text messages that describe what the user wants to achieve and see. You can, for example, insert an image of your pet and animate parts of its body, move a photo of a stream or add ripples to a snapshot featuring the sea.

Moving images aren’t the only content you can create with Lumiere. The system can replicate the style of other images, allowing the user to create something very similar, change the style of the photographs provided, insert new elements or remove unwanted ones and much more, establishing itself as highly effective tool for generating videos with AI.

The realistic and optimal result is possible thanks to the Space-Time U-Net architecture and thanks to Artificial Intelligence which, by studying the data and models to which it has access, identifies the best movement to apply to the images.

Google’s Lumiere, when the project will see the light

Currently, Google’s Lumiere it is not software distributed to the public, but it is only an experimental and research project. Despite the excellent results obtained, it will not see the light of day soon. This means that users will have to wait a long time before they can generate realistic and inconsistency-free videos from their photographs.

Before the release, Google will have to try to elimination o reduce risks of copyright infringement and those concerning security, misinformation and other important issues to evaluate when talking about generative artificial intelligence.

To know more:

  • What is video editing
  • Artificial Intelligence: what it is and what it can do for us

Leave a Reply

Your email address will not be published. Required fields are marked *